Growth of SiC by High Temperature CVD and Application of … · 2016-05-15 · Growth of SiC by High Temperature CVD and Application of Thermo-gravimetry for an In-situ Growth Rate

Growth of SiC by High Temperature CVD andApplication of Thermo-gravimetry for an In-situ

Growth Rate Measurement

von der Fakultät für Ingenieurwissenschaften, Abteilung Maschinenbau

der Universität Duisburg-Essen

zur Erlangung des akademischen Grades

DOKTOR-INGENIEUR

genehmigte Dissertation

von

Ahmed Elhaddad

aus

Ägypten

Referent:Prof. Dr. rer. nat. Burak Atakan

Korreferent: Prof. Dr. rer. nat. Markus Winterer

Tag der mündlichen Prüfung: 02.08.2010

i

ii

Abstract

Silicon Carbide (SiC) is an important compound with many benefits to man kind, rang-

ing from early usage as an abrasive to its recent use as an intrinsic semiconductor.

SiC is typically man made, since it rarely exists in nature in the form of the natural

moissanite. The production of crystalline SiC with increasing size and high quality

has been accomplished using Physical Vapor Transport (PVT) for the high power and

low frequency applications. Although high quality crystals could be produced us-

ing this method, growth defects like micropipes, dislocations, etc., could not be com-

pletely inhibited. Thus, the preparation of the SiC powder used in this process requires

additional energy, which makes the High Temperature Chemical Vapor Deposition

(HTCVD) technique more attractive and could be considered as an efficient alternative

to the PVT method, which requires lower throughput capacity than PVT due to saving

the exergy destroyed in the powder formation process (used in the PVT method) and

due to its low precursor’s cost and their disposition for continuous feeding. Therefore,

in this work, a vertical hot-wall reactor with an upward flow direction, was built and

suited for the investigation of the epitaxial growth of low defect SiC single crystalline

using the HTCVD technique. The gases are injected into the reactor through a water

cooled flange and a nozzle including an optical access for the temperature measure-

ment. The exhaust gases were removed by four openings at the top of the reactor. The

substrates were fastened on a graphite seed-holder and hanged in the reactor using a

graphite cord. The precursors used were silane (SiH4), propane (C3H8) and hydrogen

(H2) while helium was used as a carrier gas. The temperature profile was measured

by means of two color pyrometer. A maximum temperature of 2180◦C was measured

on the reactor walls, while a temperature of 2025◦C was measured on the seed-holder,

iii

which was hanged 18 cm below the outlet flange.

Non-seeded growth of polycrystalline SiC was carried out and used for indicating the

growth parameters that were later used as a reference for the setup of the epitaxial

growth experiments. In the epitaxial growth experiments the deposition was firstly

performed on on-axis 6H-SiC seeds. At a temperature of 1995◦C, a growth rate of 32

µm/h was achieved. This temperature was achieved at a substrate position of 30 cm

below the outlet flange (5 cm above the middle plane of the inductive coil). Layers

that grew on on-axis substrates were shown to have a plain surface morphology. The

growth rate has shown a significant dependency on the C3H8 within low range of its

concentrations. The step flow mechanism is activated when off-axis seed-crystals with

a tilt angle of 3.5◦ and 8◦ are used. On the film layers that were grown on the off-axis

substrate with a tilt angle of 3.5◦, flat terraces with sharp edges could be recognized

by optical microscopy. Instead, wavy surface morphology resulted on the films grown

on the 8◦ off-axis seed. Increasing the temperature beyond 1955◦C resulted in higher

growth rates on the off-axis surfaces; meanwhile, no growth rate enhancement was

obtained on the on-axis surfaces. A maximum growth rate of 100 µm/h was achieved

at a growth temperature of 2060◦C. The epitaxial growth of SiC by HTCVD was suc-

cessfully carried out for long periods up to 3 hours using the hot-wall reactor.

The growth of thick epilayers of SiC can be realized at high growth rates for several

hours, which makes it very important to measure the mass change of the substrate dur-

ing deposition. On the other hand, investigating the growth rate for a wide range of the

process parameters, can be accomplished by the application of an in-situ measuring

iv

technique, which can save a lot of experimental time and cost. Accordingly, a mag-

netic suspension balance (MSB) was successfully integrated into the hot-wall reactor

and used for the in-situ measurement of the mass change during deposition. In order

to minimize the experimental cost, this technique was only applied during non-seeded

growth experiments, where polycrystalline SiC was deposited directly on a graphite

seed-holder with 50 mm diameter. The mass change could be successfully recorded

at a growth temperature of 1950◦C, flow velocity of 0.0075 m/s and a pressure of 800

mbar. The dependency of the growth rate on the precursor (SiH4, C3H8 and H2) con-

centrations was investigated individually while the mass change was recorded in-situ.

v

********************************************

Dedication

TO

My parents

My wife

vi

Acknowledgment

I want to express my sincere gratitude to my supervisor Prof. Dr. Burak Atakan. His

kind, informative and encouraging supervision were always with me during the period

i spent on my research. He always gave me time and answered my questions with

great patience. The long discussions with him made it possible for me to understand

and realize the research tasks.

I would like to thank Prof. Dr. Winterer for his tricky suggestions and friendly support.

Great thank goes to Dr. Ulf Bergmann for his continuous support during the whole

research period and especially, for his help to write this dissertation.

I would like to express my gratitude to all those who gave his time and support to build

the hot-wall reactor. I sincerely thank Dipl.Phys. Erdal Akieldiz for his appreciated

support.

Finally I should not forget to give my special thanks to my wife Dorra whose patience

and love enabled me to complete this work.

Duisburg, Feb. 2010

Ahmed Elhaddad

vii

Contents

1 Introduction 1

2 Structure and Growth of SiC 7

2.1 Crystal Structure of SiC . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 SiC Growth Techniques . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.2.1 Growth from Melt . . . . . . . . . . . . . . . . . . . . . . . 11

2.2.2 Lely Growth . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.2.3 Seeded Sublimation Growth . . . . . . . . . . . . . . . . . . 13

2.2.4 CVD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.3 SiC Growth by High Temperature CVD . . . . . . . . . . . . . . . . 17

2.3.1 Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.3.2 Diffusion Limited Deposition . . . . . . . . . . . . . . . . . 22

2.3.3 SiC Homoepitaxial Growth . . . . . . . . . . . . . . . . . . . 24

3 Analytical Methods 27

3.1 Optical Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

3.2 Scanning Electron Microscopy (SEM) . . . . . . . . . . . . . . . . . 28

3.3 X-ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.4 Energy Dispersive X-ray Spectroscopy (EDX) . . . . . . . . . . . . . 31

3.5 In-situ Analysis of Mass Change . . . . . . . . . . . . . . . . . . . . 32

3.5.1 The Magnetic Suspension Balance . . . . . . . . . . . . . . . 34

3.5.2 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . 35

4 HTCVD System Design and Setup 41

viii

CONTENTS CONTENTS

4.1 HTCVD system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.1.1 The Hot-wall Reactor . . . . . . . . . . . . . . . . . . . . . . 43

4.2 Gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4.2.1 C-Precursor: Propane . . . . . . . . . . . . . . . . . . . . . . 47

4.2.2 Si-Precursor: Silane . . . . . . . . . . . . . . . . . . . . . . 47

4.2.3 Carrier Gas . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.3 Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

4.3.1 Seed-holder . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

4.3.2 Substrate Polytype . . . . . . . . . . . . . . . . . . . . . . . 55

4.3.3 Seed Adhering . . . . . . . . . . . . . . . . . . . . . . . . . 56

4.4 Deposition Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

4.4.1 Temperature Profile . . . . . . . . . . . . . . . . . . . . . . . 56

4.4.2 Geometrical Setup . . . . . . . . . . . . . . . . . . . . . . . 59

4.5 Experimental Procedure . . . . . . . . . . . . . . . . . . . . . . . . . 60

5 Deposition of SiC 63

5.1 Non-seeded Growth of SiC . . . . . . . . . . . . . . . . . . . . . . . 63

5.1.1 Observation of SiC Growth on Graphite Stripes . . . . . . . . 63

5.1.2 Stagnation Flow Geometry . . . . . . . . . . . . . . . . . . . 66

5.1.3 Fastening of the Seed-crystal . . . . . . . . . . . . . . . . . . 67

5.2 Seeded Growth of SiC . . . . . . . . . . . . . . . . . . . . . . . . . 69

5.2.1 Substrate Surface Treatment . . . . . . . . . . . . . . . . . . 71

5.2.2 Indication of Optimum Growth Temperature . . . . . . . . . 73

5.2.3 Improved Flow Geometry . . . . . . . . . . . . . . . . . . . 75

5.2.4 Seed-holder Improvement . . . . . . . . . . . . . . . . . . . 78

5.2.5 Improvement of Growth Conditions for Epitaxy . . . . . . . . 81

5.2.5.1 Effect of Propane Concentration on Growth Rate . . 87

5.2.6 Growth Morphology on On/Off-oriented Surfaces . . . . . . . 89

5.2.7 Effect of Silane Concentration on Growth Rate. . . . . . . . . 93

5.2.8 Dependence of Growth Rate on Temperature . . . . . . . . . 94

5.2.9 Deposition of Thick SiC Film . . . . . . . . . . . . . . . . . 99

ix

CONTENTS CONTENTS

5.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

6 In-situ Growth Rate Measurement 105

6.1 In-situ Growth Rate Measurement of Polycrystalline SiC . . . . . . . 105

6.2 Effect of Silane Concentration on Growth Rate . . . . . . . . . . . . 107

6.3 Effect of Propane Concentration on Growth Rate . . . . . . . . . . . 108

6.4 Effect of Hydrogen Concentration on Growth Rate . . . . . . . . . . 109

6.5 Vertical Placement of the Substrate . . . . . . . . . . . . . . . . . . . 110

6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

7 Conclusions 113

8 Appendix a: Technical Drawings 118

9 Appendix b: Pictures of the Hot-wall Reactor 135

Bibliography 138

x

Chapter 1

Introduction

Most semiconductor applications today are using silicon based devices. According to

the advanced process technology and development techniques, silicon crystals of high

purity and crystalline quality can be produced in mass production. However, silicon

has many limitations due to its physical properties: silicon is limited to a maximum

operation temperature of 150◦C and does not resist very high voltages. Therefore,

SiC could be the better choice according to its supreme physical and electrical prop-

erties making it very promising for the next generation of semiconductors for extreme

conditions, where their application in modern electrical engineering is essential. SiC

has many remarkable properties which make it a very promising semiconductor ma-

terial. Some of the potential applications of silicon carbide are in high-temperature,

-frequency, and -power electronic devices. Others make use of the wide bandgap:

UV radiation detectors and even blue-light lasers. Light emitting diodes (LEDs) have

already been in commercial production for some years. Also some other electronic

devices may become commercial in the near future. The large-scale manufacturing of

1

Chapter 1: Introduction CONTENTS

electronic devices requires a continuous production of good-quality wafers. In silicon

carbide growth, there are still some basic problems to be resolved that limit the com-

mercial utilization of the material. These problems are related to crystal size and both

macroscopic and microscopic defects.

As reported in the PhD thesis of Peter Råback in [1], SiC is the only stable compound

in the Si-C equilibrium system at atmospheric pressure. It was first observed in 1824

by Joens Berzelius. The properties and potential of the material were, of course, not

understood at that time. The growth of polycrystalline SiC with an electric melting

furnace was introduced by Eugene Acheson around 1885. He was also the first to

recognize it as a silicide of carbon and gave it the chemical formula SiC. The only

occurrence of SiC in nature is found in meteorites. Therefore, SiC cannot be mined

but must be manufactured with elaborate furnace techniques.

In its polycrystalline forms, silicon carbide has long been a well proved material in

high temperature, high-strength and abrasion resistant applications. Silicon carbide as

a semiconductor is a more recent discovery. In 1955, Jan Antony Lely proposed a new

method for growing high quality crystals which still bears his name [2]. From this point

on, the interest in SiC as an electronic material slowly began to gather momentum; the

first SiC conference was held in Boston in 1958. During the 60s and 70s SiC was

mainly studied in the former Soviet Union. The year 1978 saw a major step in the

development of SiC, the use of a seeded sublimation growth technique also known as

the modified Lely technique. This breakthrough led to the possibility for true bulk

crystal preparation. In section 2.2, a summary of the SiC growth techniques by CVD

is presented in details.

Different crystal forms of the same chemical composition are called polymorphs. Poly-

2

CONTENTS Chapter 1: Introduction

morphism commonly refers to a three dimensional change affected by either a com-

plete alteration of the crystal structure or a slight shift in bond angles. Polytypism is a

special type of polymorphism which occurs in certain close-packed structures. In this

phenomenon, two dimensions of the basic repeating unit cell remain constant for each

crystal structure (polytype), while the third dimension is a variable integral multiple of

a common unit perpendicular to the planes having the highest density (closest packing)

of atoms. Silicon carbide is very rich in polytypism, as more than 170 different one-

dimensional ordering sequences have been determined. For a theoretical treatment of

SiC polytypes, see [3, 4].

SiC belongs to a class of materials commonly referred to as wide-bandgap semicon-

ductors. This means that the energy gap between the valence and conduction band is

significantly larger than in silicon. It implies, for example, that it is less probable that

thermally excited electrons would jump over the gap. Therefore, SiC devices are less

sensitive to high temperatures and should be able to operate at temperatures exceeding

500◦C. The thermal conductivity of SiC is larger than that of copper. Thus the heat

generated by the devices is efficiently removed. Also such properties as high electric

feld strength and high saturation drift velocity are important for the device technology.

Consequently devices can be made smaller and more efficient. SiC is a very hard mate-

rial. This has resulted in a wide variety of applications using polycrystalline SiC. SiC is

also chemically inert and extremely radiation resistant. It may thus be used in the most

hostile environments, for example, near nuclear reactors and in outer space. Some of

the properties of silicon carbide compared to some other semiconductors are listed in

Table 1.1. It may be noticed that silicon is inferior to SiC in many respects. Diamond

would be the ultimate semiconductor for power electronics, but problems related to its

3


use appear to be even larger than in the case of SiC. There are also some other poten-

tial wide-bandgap semiconductors that compete with SiC, for example, gallium and

aluminum nitride.

Property 3C-SiC 6H-SiC Si GaAs Diamond

Bandgap [eV] 2.2 2.9 1.1 1.4 5.5

Max. Temperature[◦C] 873 1240 300 460 1100

Melting point [◦C] 1800 1800 1420 1240 ?

Physical stability excellent excellent good fair very good

Thermal conductivity [W/cm.◦C] 5.0 5.0 1.5 0.5 20

Sat. velocity [107*cm/s] 2.5 2.0 1.0 2.0 2.7

Table 1.1: Properties of silicon carbide compared to other some semiconductor materials.

GaAs stands for gallium arsenide.

The use of SiC powder as a source material for the Lely method, which was improved

to the Physical vapor transport (PVT) method using the seeded growth technique, de-

mands further improvements to the reactor technology in order to achieve long depo-

sition periods, wherein continuous feed of the source material is relatively difficult.

Thus, the PVT method requires very high energy for the sublimation of the SiC pow-

der at very high temperatures up to 2600◦C. The resulting species of the sublimation

process are transported in the gas-phase due to temperature gradients to nucleate on the

surface of the seed-crystal. In high temperature chemical vapor deposition (HTCVD),

the resulting species of the direct pyrolysis of the silane and one hydrocarbon (e.g.

propane) in the gas-phase, are used here as an input for the growth of SiC. Due to the

4

CONTENTS Chapter 1: Introduction

use of this method, the exergy lost in forming the SiC powder can be saved. Accord-

ingly, HTCVD can be considered as an energy efficient alternative for the PVT method.

Another advantages of the HTCVD lie in its ability to the use of low cost volatile pre-

cursors and their disposition by continuous feed. Although high growth rates up to 1

mm/h could be reached using the HTCVD technique, solving growth defect problems

like dislocations and the growth of hole cores (micropipes) are still requiring further

research on this technique. According to this, a BWI project researching the growth

characteristics of mono-crystalline SiC with high quality and growth rates was inspired

and started at the University of Duisburg-Essen. The tasks of this project are divided

into three parts. First, a part strives to design and build high temperature reactor to

investigate the growth process of mono-crystalline SiC by the HTCVD technique, and

the in-situ growth rate measurement via the integration of magnetic suspension balance

(MSB) into the HTCVD reactor. The results attained in this part are presented in this

work. Second comes the investigation of the crystal homogeneity by means of pho-

toluminescence at the institute of experimental physics at the Technical University of

the Bergakademie Freiberg. Finally, the particle generation in the gas-phase is being

studied, by means of mass spectroscopy in the department of Nano-Particle Process

Technology group University of Duisburg-Essen.

The dissertation is organized into six chapters. In the next chapter, a brief description

of the SiC growth techniques, the HTCVD method and the techniques applied for

the film analysis is presented. In chapter 3, the design of our HTCVD system, the

experimental setup and procedure are discussed in detail. Chapter 4 is divided into two

main parts. First, the observation of non-seeded growth of SiC on graphite substrates

5


was targeted to indicate the parameters setup required by the second part, where seeded

growth of SiC was epitaxially aimed. The grown crystals were then analyzed by XRD,

SEM and optical microscopy. In chapter 5, the in-situ growth rate measurement using

a gravimetric technique is presented. The growth rate measurement was carried out for

films grown by non-seeded HTCVD of polycrystalline SiC on a graphite seed-holder.

6

Chapter 2

Structure and Growth of SiC

2.1 Crystal Structure of SiC

Figure 2.1: Si and C atoms arranged in a tetrahedron, which is the smallest building block of

the crystal structure, as found in [5].

A brief description of SiC and its structure was introduced in the PhD thesis of Hina

Ashraf in [5]. SiC has strong chemical bonds with a short bond length (1.89 Å) be-

tween Si and C atoms. The slight difference in electro-negativity between these two

atoms gives 12 % ionicity to the otherwise covalent bonding, with the Si atom slightly

7

2.1 Crystal Structure of SiC CONTENTS

positively charged. The basic building block of the crystal is a tetrahedron consisting

of a C (Si) atom in the middle and four Si (C) atoms at the four corners, as seen in

Figure 2.1.

An important property of SiC is polytypism. Thus we can say that SiC is not a single

semiconductor but a family of semiconductors. There are many polytypes of SiC but

the simplest ones can be considered as natural super lattices. The Si-C double layer in

the stacking order ABCABC leads to the zincblende structure (3C-SiC) as discussed

in [6], while the stacking order ABAB leads to the wurtzite crystal structure (2H-SiC),

as shown in Figure 2.2. Additionally, there is a huge number of hexagonal and rhom-

bohedral polytypes with a higher complexity of their stacking order. The prevalent

occurring polytype of SiC is 6H-SiC.

Figure 2.2: Left: Wurzit Structure, Right: Zincblende Structure as found in [6].

The freedom to choose between two different positions of the second layer and by

creating an ordering in the stacking sequence of the layers, gives rise to a variety of

different polytypes. The stacking of double layers is most conveniently viewed in a

hexagonal system, as shown in Figure 2.3, with three different position of the atom

pair labeled A, B and C. The c-axis is perpendicular to the basal plane, which lies in

the plane of the close packed double layer. The three most common and important

8

CONTENTS 2.1 Crystal Structure of SiC

polytypes of SiC are 3C, 4H and 6H, although 15R and 21R are also fairly common.

Here, the Ramsdell notation is used, where the number represents the number of bi-

layers per unit cell and the letter represents the type of Bravais lattice, i.e. H stands for

hexagonal, C for cubic and R for rhombohedral. Consequently, there is no difference

between the polytypes within the basal plane. It is the stacking sequence of double

layers along the c- axis that gives rise to different polytypes.

Figure 2.3: The hexagonal system to describe different polytypes, as in [5], and the three

different positions, A, B and C of the double layers, respectively.

If the stacking sequence of the different polytypes is projected in the⟨1120

⟩plane as

indicated in Figure 2.4 and 2.5, we can observe difference in the local environment for

different atomic sites. In the turning point (is a hexagonal local lattice which repeats

its self within one structure of the polytype, in 4H and 6H there is one turning point),

the local environment is hexagonal (h) and between the turning points, the local envi-

ronment is cubic (k). 3C polytype has a cubic structure since there is no turning point,

while the 4H polytype has one cubic and one hexagonal site (h,k). The 6H polytype

has one hexagonal and two cubic sites (h,kB1B,kB2B). For the cubic and the hexag-

onal lattice site in 4H and 6H polytypes, the arrangement of the surrounding atoms

differs from the second neighbors while the two cubic lattice sites kB1B and kB2B

9

2.1 Crystal Structure of SiC CONTENTS

Figure 2.4: One stacking period of three common polytypes 3C, 4H and 6H-SiC as in [5].

10

CONTENTS 2.2 SiC Growth Techniques

differ firstly in the third neighbors.

Figure 2.5: The⟨1120

⟩plane of the three polytypes 3C, 4H and 6H, as in [5].

Cubic silicon carbide which is known as 3C-SiC, is also called beta SiC, while hexago-

nal SiC is called as alpha-SiC. The complete name of SiC polytype consists of a name

(SiC), minus sign and a suffix. The suffix consists of a number and a capital letter.

The number describes the stacking sequence and the letter the crystal system (cubic or

hexagonal)[7].

2.2 SiC Growth Techniques

This section discusses mono-crystalline SiC growth techniques except HTCVD,

which is discussed in details in section 2.3.

2.2.1 Growth from Melt

Most commercially utilized single crystal semiconductor boules are grown from a melt

or solution, but this is not a feasible option for SiC growth. SiC does not have any

liquid phase at normal engineering conditions. Calculations have indicated that stoi-

chiometric melting is possible only under pressures exceeding 105 bar at temperatures

11

2.2 SiC Growth Techniques CONTENTS

higher than 3200◦C as mentioned in [1]. Even if the solubility of carbon in silicon melt

ranges from 0.01% to 19% in the temperature interval from 1412 to 2830◦C, at high

temperatures the evaporation of silicon makes the growth unstable. The solubility of

carbon can be increased by adding certain metals to the melt (e.g., praseodymium, ter-

bium, scandium). This would, in principle, enable the use of crystal pulling techniques,

such as Czochralski growth. Unfortunately there is no crucible material available that

would be stable with these melts. It is also speculated that the solubility of the added

metals in the growing crystal is too high to be acceptable in semiconductor materials

[8, 9]. In spite of all the problems, SiC was grown from melt at 2200◦C and 150 bar

in a recent study. The crucible was made of graphite and it also acted as the carbon

source. The technology is very expensive and seems to be economically not feasible.

However, growing from a solution would avoid many of the problems related to the

growth techniques from the gas-phase.

2.2.2 Lely Growth

The Lely growth method is used even nowadays to grow the crystals of the highest

quality. The headstone of this method was placed in 1955 by the physicist Lely [2]. He

used a crucible and porous cylinder of graphite to grow his first mono-crystal of SiC by

heating SiC powder between the porous cylinder and the crucible at a temperature of

2500◦C. Increasing the crucible temperature leads to vaporization of the SiC powder

into gaseous species that flow through the cooler porous graphite. The problem he

faced was to control the pressure and the temperature. A schematic Lely geometry is

presented in Figure 2.6. In CVD the growth is driven by the initial gas composition,

whereas in the Lely method the growth is controlled by temperature gradients within

the system and by the chemical potential of the gas-phase species. The system is

12


close to chemical equilibrium, the SiC forming species have high partial pressures

and low chemical potentials due to the applied high temperatures. This leads to a

pressure gradient that results in mass transport from the hot parts to the cooler parts

of the crucible [10, 11, 12, 13, 14]. In the Lely growth, the temperature distribution

is such that in the cylindrical crucible the temperature minimum is at the inner side

of the porous graphite cylinder. Therefore the gases travel towards its inner walls.

The porous graphite holding the source provides nucleation centers for initially small

seed crystals that eventually grow larger and usually obtain an energetically favorable

hexagonal form. Unfortunately, the Lely grown crystals are limited and random in

size, in average they are about the size of a nail (3-10 mm long). Because the Lely

growth method does not require any seed crystals, it is the method from which all the

other SiC crystals originate from. The resulting crystals have low defect and micropipe

densities.

2.2.3 Seeded Sublimation Growth

The seeded sublimation growth, also known as physical vapor transport (PVT), is his-

torically referred to as the modified Lely method, reported by Tairov and Tsvetkov

[11] in 1978. PVT is a sublimation deposition process where the growth rate is pro-

portional to, and the crystallization process is facilitated by, the supersaturation of the

vapor phase. This method was further refined by Carter [15] and Stein [16] for pro-

ducing larger SiC boules. The geometry of a PVT reactor was initially quite similar

to the Lely geometry but the difference is the use of a seed crystal which results in a

more controlled nucleation, see Figure 2.7.

The seeded sublimation process is nowadays the standard method for growing bulk

13


Figure 2.6: A schematic drawing of the Lely growth method as

found in [1].

14


Figure 2.7: (a) Modified Lely Method (PVT), (b) Modified (PVT) method, as found

in [1].

mono-crystalline silicon carbide [8, 17, 18, 19]. In this process polycrystalline SiC

at the source sublimes at a high temperature, (1800-2600◦C) and low pressure. The

resulting gases travel through natural transport mechanisms (caused by natural con-

vection, where the fluid motion is not generated by any external source like a pump,

fan or suction device, etc, but only by density differences in the fluid occurring due

to temperature gradients) to the cooler seed crystal where crystallization due to super-

saturation takes place. The seed crystal is usually situated at the top of the crucible in

order to prevent contamination by falling particles. Usually, the seed is produced by

the Lely method or taken from previous growth, which was already produced by the

PVT method.

Although the seeded sublimation growth is the most promising method for producing

SiC crystals, and has been known for more than twenty five years, there are still some

major difficulties involved in the process. The polytype formation and growth shape

are poorly controlled and the doping is nonuniform. There are also severe defects, such

15


as micropipes and dislocations in the grown crystals.

2.2.4 CVD

Chemical vapor deposition is a widely used method for depositing thin or thick films

with high quality and well defined chemical composition and structural uniformity. In

a typical chemical vapor deposition process the substrate is exposed to one or more

volatile precursors, which react and/or decompose on the substrate surface to produce

the desired deposit. The activation energy for the reaction can be overcome by various

methods, where the most common approach is to heat the substrate. Volatile byprod-

ucts are generally also produced, which are removed by the gas flow through the reac-

tion chamber. The main benefit derived by a chemical vapor deposition process is the

resulting uniform, adherent and reproducible films. Often the main disadvantages lie

in the need to resort to dangerous and toxic chemicals to obtain the desired deposition

along with the high temperatures necessary for some of the reactions. CVD technology

opens possibilities to prepare new materials and structures for various applications for

many industrial products.

Using this method, the direct growth of SiC from the gas-phase was achieved epitax-

ially at a temperature range of 1200-1700◦C in a hot or cold wall reactor [20, 21, 22,

23, 24]. By the injection of SiH4 and hydrocarbon diluted in carrier gas (Ar, He or H2)

into the reactor, the deposition of several micrometers of mono-crystalline SiC can be

accomplished. By increasing the temperature, the growth rate increases, but currently

problems related to the growth control , like temperature, pressure and the concentra-

tion of the gas-phase species, become more severe, and problems such as homogeneous

nucleation in the gas phase may occur. These problems might be overcome by a very

16

CONTENTS 2.3 SiC Growth by High Temperature CVD

careful control of the thermodynamic conditions.

However, the growth rate in the cold-wall reactor is limited to 3-5 µm/h [21, 23], while

in the hot-wall reactor, the highest reported growth rate lies around 10 µm/h [22].

Vertical reactors have also been investigated. Both, rotating disk reactors [25] and

a multi-wafer rotating planetary reactors [26], have been developed to improve layer

uniformity and reactor throughput. Fast epitaxial growth using this technique was

reached by Danielsson [20], who introduced the growth of 4H-SiC with high growth

rates of 10-49 µm/h. Later on, two other groups also achieved growth rates above 10

µm/h using the vertical up-flow configuration in a chimney [27] or a radiation heated

CVD reactor [28]. Fast epitaxial growth of high quality 4H-SiC in vertical hot-wall

CVD reactor at 25-60 µm/h has been also reached by Fujihira [29].

A higher growth rate than 10 µm/h was investigated in a vertical hot-wall reactor.

By the use of increased temperatures (1650-1850◦C) and high reactant concentrations,

this process is shown to enable growth rates up to 50 µm/h and demonstrates a material

quality comparable to established CVD techniques until growth rates of 25 µm/h [30].

Generally, the growth rates in CVD are too low to allow boule production. Usually tens

of micrometers an hour are attainable, which makes this method not very attractive for

the application where single crystalline SiC is required to be produced at high growth

rates.

2.3 SiC Growth by High Temperature CVD

The first successfully grown SiC crystal by HTCVD was introduced 1996 at the De-

partment of Physics and Measurement Technology of Linköping University in Sweden

[31]. They have constructed a prototype of a vertical hot wall reactor with relatively

17

2.3 SiC Growth by High Temperature CVD CONTENTS

small inner diameter (30 mm) of the crucible, see Figure 2.8. The reactor design based

on the technology used in the modified PVT reactor. The basic difference lies in the

use of an inlet duct at the bottom of the reactor to inject the process gases instead of

using SiC powder as usual in the PVT method. The crucible is covered by a graphite

lid with four small holes for the gas removal, while the substrate was glued on the

center of the lid. The crucible is closed from the bottom by a graphite plate with one

hole for the gas inlet. The temperature measurements are accomplished by a two-color

pyrometer focused on the lid of the crucible. The crucible is inductively heated by a

50 kW middle frequency generator operating. All sides of the crucible are enclosed

by graphite foam for thermal insulation and the assembly is placed on a quartz support

inside an air-cooled quartz tube. H2 and Ar or a mixture of both gases was chosen as a

carrier gas. SiH4 was used as a source for Si species, while small doses of C3H8 was

carefully added to the mixture with small amounts, since at very high temperatures the

decomposition rate of the crucible is further increased developing more hydrocarbons

into the gas-phase. A maximum growth rate of 300 µm/h was reached at a temperature

of 2300◦C by this reactor. In 2004, a 2-inch, 15 mm thick SiC single crystal grown by

HTCVD has been commercially introduced by Okmetic AB [32]. The 4H-SiC crystal

demonstrated a micropipe density down to 0.3 cm−2.

In this process, the precursors or the gas mixture is injected into the crucible through

the bottom inlet. The higher the temperature the larger is the probability that the re-

actants will be thermally pyrolyzed producing species or radicals that will attach to

the surface thus leading to epitaxial growth [1]. When the substrate temperature Ts

is increased the probability of sticking increases, but also the etch rate from the sur-

face increases. The growth rate is therefore determined by the desorption of the re-

18


Figure 2.8: Reactor prototype used for the HTCVD of crystalline SiC. The reac-

tor was introduced by the Department of Physics and Measurement Technology of

Linköping University in Sweden in 1996 [31].

19


action products and by the etch rate of the surface, and by the diffusion dominated

mass transport of the source molecules. The sticking coefficient Sc depends on both Ji

the impingement flux rate and the rate constants due to this relation Sc = Rr/Ji [33],

where Rr is the desorption rate. Rr is dependent on the global activation energy and

the temperature. If the activation energy is high enough (Edesorption-Eadsorption>0) the

temperature must be increased to make the exponential term (Er-Ed)/RTs small. On

other hand, the resorption rate Er< Ed will be decreased by increasing the surface tem-

perature Ts. Inherent advantages of this technique include continuous supply of the

source material, the relatively economical availability of high purity gases, the control

of the C/Si ratio and the ability of pulling the growing crystal. Thus, the use of SiC

powder in the Lely and PVT methods demands much energy, where high amount of

the gases potential exergy are lost in the formation of the SiC powder. In HTCVD, the

direct pyrolysis of the precursors into the process saves the exergy lost for the powder

formation. That’s why, the HTCVD method is considered as more energy efficient

than both other methods of Lely and PVT.

2.3.1 Transport

The gross reaction equation for the HTCVD of SiC using both precursors, silane and

propane, can be given due to the following reversible reaction:

3 SiH4 + C3H8⇔ 3 SiC + 10 H2

The HTCVD process with silane and propane precusors is described in Figure 2.9 and

explained particularly in the following four steps.

• The process gases (typically silane and a hydrocarbon diluted in a helium carrier

gas flow) are transported upwards to a substrate or seed-crystal fixed on a seed-

20


Figure 2.9: A schematic description of the HTCVD process with SiH4 and C3H8 as precur-

sors.

21


holder in a hot-wall reactor as presented in section 4.1.1.

• Gas-phase chemical reactions are initiated as the precursors flow through the

hot zone of the reactor, typically made of graphite, which together with the

design of the seed-holder will determine the growth species concentration and

temperature-gradient profile near the seed surface. In this step, the temperature

must be maintained very high so that Si and the hydrocarbon radicals do not

react together and build stable SiC clusters, which normally depletes and fall in

the opposite direction of the flow stream.

• The vapor species or radicals released by the hot zone are transported by flow

system determined by the reactor design to the seed, which is maintained at lower

temperature than the hot zone. Thus, the high temperature established in the

hot-zone causes a reduction of the species chemical potential, which results in

increased concentrations of the major species ensuring high supersaturation and

growth rate. This and the chemical potential of the vapor species and radicals

ensure high supersaturation and growth rate.

• Removal of by-product gases from the CVD chamber through the exhaust sys-

tem.

2.3.2 Diffusion Limited Deposition

As previously explained in section 2.3, high temperature and precursor concentrations

are required when fast epitaxial growth of crystalline SiC is desired. The influence

of the temperature on the growth kinetics can be studied on the basis of an Arrhe-

nius diagram, as shown in Figure 2.10, which presents three zones wherein the species

22


transport to the seed surface is controlled due to different mechanisms. In the low

temperature region, the growth rate can be enhanced by increasing the temperature.

This indicates a kinetically controlled process, where increased growth temperatures

enhance the decomposition of SixCy clusters and thus the diffusion to seed surface,

which consequentially results in an enhanced supersaturation at the growing SiC sur-

face. In the middle range where higher growth temperatures are applied, the growth

rate follows a smoother dependence, which may be more ascribed to a mass-transport

limited regime. The growth rate is then generally influenced by the temperature gradi-

ent applied to the seed crystal. In the highest temperature region, for a given temper-

ature gradient, a further increase of the growth temperature decreases the net growth

rate, which indicates the onset of equilibrium limit or the onset of film re-evaporation.

Figure 2.10: A schematic description of the growth rate dependence on the temperature as

discussed in [33]. In this diagram, three zones of the growth process are presented, which are

individually explained in the text in reference to the HTCVD process of mono- crystalline SiC.

The work introduced by Ellison [34] in 1999 observed that the growth process is lim-

23


ited by gas-phase chemistry within a seed temperature range of 2160-2230◦C, where

the growth rate is shown to be enhanceable by increasing the temperature in this range.

Further increase of the temperature beyond 2230◦C resulted in constant growth rate,

which possibly indicates a growth mechanism limited by mass transport. In this case,

increasing the precursor concentrations is usually favorable to increase the growth rate.

2.3.3 SiC Homoepitaxial Growth

The plane orientation of the seeds surface plays an important role in determining the

SiC polytype that will grow. For example, a parallel plane to the crystalline layers (A

or B) is normally coded by the number (0001) and usually called well oriented surface,

which is known as basal plane. A misaligned surface to the [0001] direction with a

small tilt angle to the well oriented surface develops a so-called off-axis surface. The

growth mechanism on both surface planes is schematically depicted in Figure 2.11.

Homoepitaxial growth of SiC is basically determined by the polytype of the seed and

its surface plane. Under both conditions, homoepitaxial growth of SiC can be ac-

complished by step controlled epitaxy [23, 35, 36]. Step controlled epitaxy is based

upon the growing epilayers on, usually, 3.5◦ or 8◦ off-axis surfaces, resulting in a sur-

face with atomic steps and flat terraces between steps. When growth conditions are

properly controlled and there is a sufficiently short distance between steps, Si and C

adatoms impinging onto the growth surface find their way to steps where they bond

and incorporate into the crystal. Thus, ordered lateral (step flow) growth takes place,

which enables the polytypic stacking sequence of the substrate to be exactly mirrored

in the growing epilayer. When growth conditions are not properly controlled or when

steps are too far apart (as can occur with SiC substrate surfaces that are polished at

24


Figure 2.11: a) On-axis surface: also called well-oriented. On these surfaces, the growth

mechanism will perform on terrace area via two-dimensional nucleation, expanding laterally.

b) Off-axis surface: the orientation of seed surface is misaligned to [0001] direction with a tilt

angle of 3.5◦ or 8◦, which develops high density of steps and narrowing terraces widths, which

allows the replication of the epilayer structure. This mechanism is known as step-flow epitaxy

[5].

25


a tilt angle less than 1◦ to the basal plane), growth adatoms can nucleate and bond in

the middle of terraces instead of at the steps, which leads to heteroepitaxial growth

of poor-quality 3C-SiC [23, 26]. To help prevent 2D-nucleation of 3C-SiC (triangular

inclusions) during epitaxial growth, most commercial 4H- and 6H-SiC substrates are

polished to tilt angles of 8◦ and3.5◦ off the (0001) basal plane, respectively.

26

Chapter 3

Analytical Methods

The deposited film has to be analyzed in order to identify its physical and chemical

properties where these properties are important in industrial applications. Morphol-

ogy, phase composition and chemical composition are investigated in this study using

techniques that include X-ray diffraction (XRD), microscopy, scanning electron mi-

croscopy (SEM) and energy dispersive X-ray spectroscopy (EDX).

3.1 Optical Microscopy

Microscopy is the technical field of using microscopes to view samples or objects.

There are three well-known branches of microscopy, optical, electron and scanning

probe microscopy. Optical and electron microscopy involve the diffraction, reflection,

or refraction of electromagnetic radiation/electron beam interacting with the subject of

study, and the subsequent collection of this scattered radiation in order to build up an

image. This process may be carried out by wide-field irradiation of the sample (for

example standard light microscopy and transmission electron microscopy) or by scan-

ning of a fine beam over the sample (for example confocal laser scanning microscopy

27

3.2 Scanning Electron Microscopy (SEM) CONTENTS

and scanning electron microscopy). Scanning probe microscopy involves the interac-

tion of a scanning probe with the surface or object of interest. The development of

microscopy revolutionized biology and remains an essential tool in that science, along

with many others including materials science and numerous engineering disciplines.

Optical or light microscopy involves passing visible light transmitted through or re-

flected from the sample through a single or multiple lenses to allow a magnified view

of the sample. The resulting image can be detected directly by the eye, imaged on a

photographic plate or captured digitally. The the system of lenses and imaging equip-

ment, along with the appropriate lighting equipment, sample stage and support, makes

up the basic light microscope. The optical microscope, often referred to as the "light

microscope", is a type of microscope which uses visible light and a system of lenses to

magnify images of small samples. Optical microscopes are the oldest and simplest of

the microscopes. Digital microscopes are now available which use a CCD camera to

examine a sample, and the image is shown directly on a computer screen without the

need for expensive optics such as eye-pieces. Other microscopic methods which do

not use visible light include scanning electron microscopy and transmission electron

microscopy.

3.2 Scanning Electron Microscopy (SEM)

The surfaces morphologies of the films were analyzed by Scanning Electron Mi-

croscopy (SEM). The SEM is a microscope that uses electrons rather than light to

form an image. There are many advantages of using the SEM instead of a light mi-

croscope according to its ease of sample observation, higher magnification and higher

resolution. The principle of the SEM is to focus a beam of primary electrons onto a

28

CONTENTS 3.3 X-ray Diffraction

sample, and to collect secondary electrons scattered from the sample. An image is cre-

ated by scanning the sample surface point by point by a focused beam of electrons and

to reconstruct the image from the scattered intensities. The sample is placed inside a

vacuum chamber. After the chamber is evacuated, an electron gun emits a beam of high

energy electrons. This beam travels down ward through a series of magnetic lenses de-

signed to focus the electrons to a very fine spot. A set of scanning coils moves the

focused beam back and forth across the sample, row by row. As the electron beam hits

each spot on the sample, secondary electrons are scattered from its surface. A detector

counts these electrons and sends the signals to an amplifier. The final image is built up

from the number of electrons emitted from each spot on the sample. A scanning elec-

tron microscope purchased from FEI company was applied for analyzing the surface

morphology of our grown crystals. The microscope has a type of 400 FEG Quanta. It

has a magnification capacity of 7x - 1000,000 in high and ultra high vacuum. It has a

resolution of 2-3.5 keV nm.

3.3 X-ray Diffraction

The films may be deposited in amorphous or crystalline phases. The atoms of a crys-

talline film are arranged in a regular pattern while the atoms of an amorphous film are

arranged in a random way. The arranged atoms of a crystal form a series of paral-

lel lattice planes separated from one another by a distance d, which varies according

to the nature of the material. For any crystal, planes exist in a number of different

orientations each with its own specific distance d. X-ray Diffraction (XRD) allows

the identification of the phase composition of the analyzed films if they are crystalline

or polycrystalline. X-ray diffractometers consist of an X-ray generator, a goniometer

29

3.3 X-ray Diffraction CONTENTS

(angle-measuring device), a sample holder, and an X-ray detector. X-rays can be gen-

erated within a sealed tube under vacuum. A tungsten filament is fixed within the tube

and connected to high voltage transformer. When a voltage is applied to the filament

at a current of 3 A, electrons are emitted and rapidly drawn to the target, often made

of copper (λ= 1.5418 A). When these electrons hit the target, X-rays are produced.

The emitted wavelength is characteristic for elements of that target. These X-rays are

collimated and directed onto the substrate. The X-ray beam hits the sample and the

detector records the X-ray intensity diffracted at the substrate. The distances between

the adjacent lattice planes can be calculated by applying Bragg’s Law, see Figure 3.1.

nλ = 2d ∗ sin θ (3.1)

Figure 3.1: Bragg’s Law.

where n is the order of diffraction (0,1,2,3,...), λ is the wavelength of the incident X-ray

beam, d is the distance between adjacent lattice planes, and θ is the angle of incidence

of the X-ray beam. The diffraction angle 2θ is equal to twice the incident angle θ.

The goniometer is motorized and moves through a range of the angle 2θ. Each time

the Bragg condition is satisfied, the detector measures the intensity of the reflected

30

CONTENTS 3.4 Energy Dispersive X-ray Spectroscopy (EDX)

radiation. An X-ray detector records the diffracted beam intensity as a function of

the angle (2θ). Every crystalline material will give a characteristic diffraction pattern,

which can be used for the identification of its structure. The plot of XRD patterns used

to identify the type of material by comparing them with standard XRD patterns in a

database. When XRD is applied for the characterization of powder or polycrystalline

samples, each particle of the powder is a tiny crystal, or assemblage of smaller crystals,

oriented randomly with respect to the incident beam. The result is that every set of the

powder or the polycrystalline lattice planes can be capable of diffraction, but just in

the case that its orientation is making a correct Bragg angle. In the XRD of single

crystalline, the rotating crystal method is used, where the crystal is rotated about only

one axis, which will lead to that only a particular set of lattice planes will, for an

instance, make the correct Bragg angle for diffieration of the monochromatic incident

beam and that instant a diffracted beam will be formed.

3.4 Energy Dispersive X-ray Spectroscopy (EDX)

Energy Dispersive X-ray spectroscopy (EDX) analysis was used for the deposited films

in order to evaluate their chemical composition. The scanning electron microscope is

also equipped by an EDX module. EDX is a micro-analytical technique that uses the

characteristic spectrum of X-rays emitted by the sample after excitation by high en-

ergy electrons to obtain information about its elemental composition. During EDX

analysis, the specimen is bombarded with an electron beam inside the scanning elec-

tron microscope. The bombarding electrons collide with the specimen atoms’ own

electrons, knocking some of them off in the process. A position vacated by an ejected

inner shell electron is eventually occupied by a higher energy electron from an outer

31

3.5 In-situ Analysis of Mass Change CONTENTS

shell. To be able to do so, however, the transferring outer electron must give up some

of its energy by emitting an X-ray photon. The number and energy of the X-ray pho-

ton emitted from a specimen can be measured by an energy dispersive spectrometer.

As the energy of the X-rays are characteristic of the difference in energy between the

two shells, and of the atomic structure of the element from which they where emitted,

this allows the elemental composition of the specimen to be measured. A quantita-

tive analysis is possible using appropriate calibration. Elements of low atomic number

such as hydrogen, which has only one electron that is impossible to be replaced by

other when it is knocked off, are difficult to detect by EDX. An EDX system coupled

to the laser electron microscope of type "Genesis 4000" was applied for analyzing our

film’s compositions.

3.5 In-situ Analysis of Mass Change

Laser reflectivity was used to measure the growth and etch rates in-situ during film

deposition systems. Laser diffraction is an optical technique, which was widely in-

troduced as convenient way to measure and detect the onset of grown layers in CVD

systems. It was applied for the in-situ measurement of growth rate during CdTe single

crystal growth from the vapour phase [37] and in plasma-enhanced chemical vapour

deposition of vertically aligned multi-wall carbon nanotube films [38]. Such an op-

tical technique is applicable when the substrates have a flat and transparent surface.

Usually, films grow randomly three dimensional patterns. The films in this case have

rough, non-flat surface morphologies that do not allow an accurate characterization

using optical methods. Therefore, as the film starts to grow accumulatively with such

a surface morphology, it becomes inaccurate to use such techniques for growth rate

32

CONTENTS 3.5 In-situ Analysis of Mass Change

measurement. In addition to this, optical techniques can not measure the growth rates

directly in µm/h or g/h, since they can only measure the growth on definite portions of

the substrates, where the laser beam is focused on, meanwhile on neighboring portions,

more or less thick layers were already grown.

Gravimetric techniques are more adequate to be applied for direct growth rate mea-

surement in CVD systems. They are applicable for relatively heavy substrates or

substrate holders. According to the physical principle of the gravimetric technique

a very stable vacuum environment is required according to its high sensitivity to the

surrounding conditions. Gravimetric techniques were used for the study of kinetics

in CVD systems, like graphite and diamond [39]. Weiner [40] and Evans [41] have

used gravimetric microbalances extensively for determining the kinetics of diamond

film growth. It was also used by Salifu [42] for the in-situ growth rate measurement

in plasma processing of opaque materials. Till present, the gravimetric technique was

not used at temperatures higher than 1300◦C, where experimental difficulties are large.

D. Neuschütz and his group have applied the microbalance in a hot-wall reactor to

measure the growth rates of Al2O3 at a temperature range of 900-1200◦C in [43].

The gravimetric microbalance has not been used previously in SiC HTCVD process.

SiC HTCVD systems are difficult to combine with the gravimetric technique because

the substrate must hang freely, the temperature is very high and the flow around the

substrate should be stable to achieve accurate measurements. To overcome these prob-

lems a magnetic suspension balance MSB was coupled to our hot-wall reactor at lam-

inar flow conditions for direct weight measurements during the growth of polycrys-

talline SiC. The work presented in this chapter focuses on the in-situ growth rate mea-

surement of polycrystalline SiC in a non-seeded HTCVD process as an initial step

33


for its future application in seeded, epitaxial growth of SiC by HTCVD. The parame-

ters used in these experiments were the same as applied in the experiments that were

presented in the previous chapter.

3.5.1 The Magnetic Suspension Balance

Figure 3.2: Schematic description to the measuring principle of the magnetic suspension bal-

ance, see the MSB manual of Rubotherm GmbH.

The Magnetic Suspension Balance (MSB) makes it possible to weigh samples contact-

lessly under nearly all environments with a balance located at ambient conditions. The

sample is located in the measuring cell and can be (un)coupled specifically from/to the

balance via a contactless magnetic suspension coupling. An electromagnet is attached

to the bottom of the balance. It lifts a so-called suspension magnet, which consists

of a permanent magnet, a sensor core and measuring load decoupling cage. The elec-

tromagnet, which is attached to the weighing hook of the weighing cell, maintains a

freely suspended state of the suspension magnet via an electronic control unit. Thereby

different vertical positions are possible. First the zero point position (ZP) in which the

34


suspension part suspends alone and contactlessly and thus represents the unburdened

balance. And second the measuring-point position (MP), in which the suspension part

reaches a higher vertical position, thereby the sample is coupled to the balance and the

balance and transmits the weight of the sample to the balance transmitting the weight

of the sample to the balance. This principle is illustrated in the following picture. A

schematic presentation of the MSB principle is shown in Figure 3.2.

3.5.2 Experimental setup

Figure 3.3: Picture of the MSB coupled with the HTCVD reactor during the growth process

of polycrystalline SiC.

The MSB was connected to the HTCVD reactor as illustrated in Figure 3.3, the tem-

perature was measured twice on the inner walls of the susceptor, firstly with a zero

flow and secondly with only a helium flow of 0.0075 m/s. A non-constant distribution

of the susceptor temperature was recognized showing a trivial temperature difference

35


between both flow conditions. A maximum temperature of 2460 K could be reached

on the inner susceptor walls. The MSB is connected to the upper flange, as seen in

Figure 3.3. An additional helium flow of 300 sccm was used for purging the MSB and

its connection elements.

As known for the growth of mono-crystalline SiC by HTCVD, seed crystals are neces-

sary for the epitaxial growth, which were usually fixed on a graphite lid or seed-holder.

The design of the hot-wall reactor deals with a seed-holder, which has a geometry of

a disc with 50 mm diameter, thickness of 8 mm and a density of 1.83 g/cm3. In the

seeded growth experiments presented in the previous chapter, the graphite disc (seed-

holder) was hanged freely inside the reactor by means of a graphite cord, which was

tied to the outlet flange, while in non-seeded growth experiments, the graphite cord was

tied to the hook of the MSB. A distance of 30 cm was measured between the bottom of

the outlet flange and the seed-holder. A growth temperature of 1950◦C was achieved

after a heating period of 30 minutes. Thereafter, both precursors SiH4 and C3H8 were

added at flow rates of 20 sccm and 150 sccm leading the concentration values of 0.132

mol/m3 and 0.994 mol/m3, respectively. Both concentrations were chosen according

to the results published by Elisson in [44], whereby it was observed that a very low

ratio of C3H8/SiH4 (nearly 1:8) is necessary for an optimal growth rate of SiC. Helium

was chosen as a carrier gas and added at 6 slm with a flow velocity of 0.0075 m/s. As

reported by B. Sundqqvist in [32], thick epitaxial layers of SiC could be successfully

grown at a pressure of approximately 1 atm. Therefore, a value of 800 mbar was used

in our experiments to avoid any danger that can be caused due to the pressure increase,

which can be caused due to a sudden breakdown of the pump or the pressure controller.

In order to check the accuracy of the MSB, the mass of the seed-holder was measured

36


Figure 3.4: The seed-holder mass (in g) is plotted versus a pressure range of 1-1000 mbar in

helium and air atmospheres in order to check the measurement’s accuracy of the MSB at room

temperature.

37


at different pressures in helium and air, individually. Figure 3.4 shows the seed-holder

mass recorded versus the pressure in both cases. At low pressures the density of the

medium is reduced, which reduces the buoyancy force and consequentially reduces

the mass recorded for the seed-holder. The difference between the measured values at

pressures of 40 mbar and 1 atm is 0.02 g in air and 0.003 g in helium. The latter value

is much less than the one measured in air, which is expected due to the lower density of

helium compared with air, which results in a lower buoyant force. Another experiment

was carried out to check the accuracy of the MSB under flow. The seed-holder mass

was measured at different flow rates as shown in Figure 3.5. The mass was firstly

measured at 800 mbar in a stationary environment. Afterwards, the flow velocity was

increased in steps by 0.0025 m/s (2 slm) every minute. The first flow acceleration from

0 slm to 2 slm had the most violent impact on the values recorded for the seed-holder

mass. Instable mass values were recorded in this period. A further increase of the flow

velocity has shown a smaller effect on the mass recorded. However, increasing the flow

velocity by increasing the flow rate every minute leads to pressure fluctuations around

the set point of 800 mbar, which leads to oscillating values of the mass recorded. Such

an instability requires a small period of time till this effect disappears.

The direct growth of polycrystalline SiC on the 2-inch seed-holder was analysed on

the basis of the mass change, which was recorded by means of the MSB. In order to

investigate the influence of the precursor concentrations on the growth rate of polycrys-

talline SiC, five experiments were carried out. The mass change was recorded in-situ

for analyzing the influence of variating the precursor concentrations on the growth rate.

• Firstly, the mass change was recorded during 45 minutes including a total de-

position period of 20 minutes. SiH4 and C3H8 were added at flow rates of 200

38


Figure 3.5: The seed-holder mass is recorded at different carrier gas flow velocities. The flow

velocity was increased by increasing the flow rate by 2 slm after a period of one minute. Fluc-

tuations of the weight values that are visible in the diagram are caused by the sudden increase

of the flow rate, which lead to a pressure disturbances and finally results in low accuracy of the

weight measurement.

39


sccm and 20 sccm in a carrier gas of helium leading to both concentrations of

1.315 mol/m3 and 0.132 mol/m3, respectively.

• Afterwards, the C3H8 concentration was held at a constant value of 0.132

mol/m3, while the SiH4 was added in gradual increase every 10 minutes at five

different concentrations of 0.668, 0.994, 1.315, 1.631, 1.941 and 2.247 mol/m3.

• In the third experiment, the SiH4 concentration was held at a constant value of

0.669 mol/m3 (100 sccm), while the C3H8 concentration was started with 0.0335

mol/m3 (5 sccm) and increased gradually to 0.067, 0.101 and 0.134 mol/m3 every

five minutes.

• In the fourth experiment, the SiH4 and C3H8 concentrations were held at a con-

stant value of 1.304 mol/m3 and 0.1304 mol/m3, meanwhile H2 was added at

different concentrations of 0, 0.326, 0.647, 0.963 and 1.274 mol/m3 that were

gradually increased every five minutes.

• Finally, the seed-holder was hanged vertically (circular area parallel to the flow

stream lines) allowing the possibility to be deposited on both of its sides. The

temperature measured on the lower side of the seed-holder was 1950◦C, the pres-

sure was held constant at 800 mbar. The SiH4 flow rate was held constantly at

200 sccm with a concentration of 1.318 mol/m3, while the flow rate of C3H8 was

initially set to zero in the first 6 minutes and then gradually increased to 5 sccm

and 10 sccm leading to both concentration values of 0.0329 mol/m3 and 0.0658

mol/m3, respectively.

40

Chapter 4

HTCVD System Design and Setup

4.1 HTCVD system

The HTCVD technique can simply be described as chemical vapor deposition CVD

at high temperatures, hence the name high temperature CVD: HTCVD. The growth

process however, differs strongly from that of the CVD process due to the significant

sublimation and etch rates at the extreme growth temperatures 1800-2300◦C. In or-

der to reach such high temperatures, inductive heating of a graphite susceptor is used.

The commonly used components in CVD systems are thus required in HTCVD. Our

HTCVD system is described in Figure 4.1, which is mainly comprised of hot-wall re-

actor (HWR)(1) with a graphite susceptor, middle frequency generator (MFG) and in-

duction coil (2), pump and pressure regulator (3), mass flow controllers (4), gas sources

(5) and pyrometer (6). The MFG is purchased from Hüttinger Elektronik GmbH and

can deliver a heating power up to 50 kW at approximately 20 kHz. The induction coil

is made of copper and water cooled. The MFCs are purchased from MKS Instruments

Germany and have different capacities in order to control the gas flow rates. The pump

41

4.1 HTCVD system CONTENTS

Figure 4.1: A schematic diagram of the HTCVD system. (1) hot-wall reactor (HWR), (2) inlet

flange, (3) nozzle, (4) outlet flange, (5) pyrometer, (6) pump and pressure controller, (7) middle

frequency generator (MFG) and induction coil, (8) optional connection for mass spectrometry

(9), (10) gases source, (11) mass flow controllers.

42

CONTENTS 4.1 HTCVD system

was purchased from Busch GmbH and has a main pumping speed of 61 m3/h. The

pyrometer is purchased from Keller HCW and is capable to measure the temperature

in a range of 1000-3000◦C. The reactor is mainly comprised of inlet flange (7), outlet

flange (8), nozzle (9), graphite susceptor, foam graphite and quartz tube. The design

of the reactor upper flange enables the optional connection of the MSB (10) or mass

spectrometer (11). Section 3.1.1 discusses the reactor design.

4.1.1 The Hot-wall Reactor

As previously introduced in section 2.3, in the HTCVD process, the precursors or the

gas mixture are injected into the susceptor in upward direction, where higher temper-

atures are established on the inner susceotor walls. This leads to the pyrolysis of the

reactants producing species or radicals that will be transported by free convection due

to thermal gradients or by forced convection to the substrate or the seed-surface, where

the growth mostly proceeds by diffusion. In order to thermally realize this, high tem-

peratures in a range of (1800-2300◦C) are required. Inductive heating has proofed a

reliable performance of the crucible or the susceptor heating in order to reach such a

high temperatures. The basic components of the induction heating are an AC power

supply, induction coil, and workpiece (sample to be heated or treated). The power

supply sends alternating current through the coil, generating a magnetic field. When

the workpiece is placed inside the coil, the magnetic field induces eddy currents in the

workpiece, generating precise amounts of localized heat without any physical contact

between the coil and the workpiece. The eddy current is an electric current circulat-

ing wholly within a mass of a conductor or a semiconductor. As known, graphite was

usually chosen for the susceptor material due to its high temperature resistance up to

2500◦C in vacuum environments that contain no oxygen. Unlike graphite, it becomes

43


critical when other materials are used in the reactor, like stainless steel, which is able

to withstand only temperatures up to 1200◦C. Applying higher temperatures on it will

increase the probability of its melting and consequentially imposing impurities into the

process, which must be avoided in HTCVD of SiC. In such case, water cooling and

a special welding design are very necessary for the reactor construction. One of the

designs of the hot-wall reactor used for HTCVD is introduced in section 2.3, Figure

2.8. This design utilizes a crucible made of graphite, which is partially closed at its

both ends. The seed-crystal is fastened on the bottom of the lid. Instead, our design

deals with a susceptor made of simple graphite cylinder and a free floating seed-holder,

whereon the seed-crystal can be fastened. This allows more flexibility to optimize the

seed-holder position for improving the growth conditions. Thus, the temperature distri-

bution on the inner side of the susceptor, seed-holder temperature (growth temperature)

and the in-situ detection of this parameter during the growth, are important for enhanc-

ing the growth conditions and increasing the growth rate. Therefore, the design of the

hot-wall reactor was carried out as a basic part of this research, which can deal with

the deposition on SiC wafers with large diameters up to 75 mm.

A side view of the reactor cross section is shown in Figure 4.2. The susceptor (1)

was made of a hard graphite tube of 130 mm inner diameter, 154 mm outer diameter

and 600 mm length. It was purchased from SG Carbon and has a density of 1.82

g/cm3. The outer side of the susceptor was covered by foam graphite (2) for thermal

insulation. Since a vertical type reactor with an upward flow direction was desired to

avoid unsymmetrical free convection, the inlet flange (3) was designed in which it is

possible to be connected with a nozzle (4) by means of an ISO CF 250 flange (5). The

precursors are fed into the reactor through four small holes (6) that are bored in the

44

CONTENTS 4.1 HTCVD system

Figure 4.2: Two dimensional view of the reactor cross section, including the inlet and outlet

flanges, susceptor, graphite foam, quartz tube, nozzle and the optical access. The detailed

description is given in the text.

45


nozzle. The inlet flange comprises of two water cooled rings (7) that are connected

to each other by welding. The flange is connected through four holes (9) with water

source, further four holes (10) were allocated for purging the foam graphite. Both parts

are inserted in each and welded at the perimeters (8). This design was intentionally

done to shift the welding butt far enough from the thermal radiation area. The duct is

water cooled at (11) and can be connected to an ISO CF 160 flange at (12). Flange (23)

has an optical access and is connected to the bottom of flange (12), which comprises

of four parts as shown in Figure 4.2. A transparent plate of pyrex glass (22) with a

diameter of 140 mm and 8 mm thickness is sealed by means of o-ring with a cross

section diameter of 5 mm to allow temperature measurement by the pyrometer. This

flange is connected to the water source through one hole at (24), whereas on the other

side, another hole was designed for the inlet of purging flow at (25). This flange was

permanently cooled by air keeping its temperature less than 300◦C.

The assembly is centered in a quartz tube (14) and sealed at its both ends by two

O-Ring and the steel parts that are scaled in (15). The outlet flange is also water

cooled and comprises of two plates (16) that are welded at their perimeters (17). For

the removal of the gases, four canals (18) were designed passing by the water cooled

area to connect the flange by the pump. An ISO CF 250 flange (19) was built on the

upper side of the flange for its connection with the magnetic suspension balance or

with the mass spectrometer. In the center of the flange, a relatively large hole (21) of

100 mm diameter was made to enable inserting the seed-holder downwards into the

susceptor(26), which is hanged by means of a graphite cord (27). The graphite tube

(28) was put on the inner side of the nozzle for separating the flow from the susceptor

walls due to an external carrier gas flow, which is fed into the reactor at (29).

46

CONTENTS 4.2 Gases

4.2 Gases

Chemical sources may be classified as: Inert gases such as argon and nitrogen, which

are used as carrier gases of a precursor vapor. Reactant gases are required such as

oxygen for oxidation or hydrogen for hydrogenation or reduction, (producing volatile

hydrocarbon compounds). Precursors are the source material of the deposits, which

should be thermally stable at room temperature and volatile at low temperatures. Mass

flow controllers are used to control and mix the amount of gases fed into the reaction

chamber during the deposition process.

4.2.1 C-Precursor: Propane

Under normal atmospheric pressure and temperature, propane is a gas. Under moderate

pressure and/or lower temperatures, however, propane changes into a liquid. Propane

is easily stored as a liquid in pressurized tanks. It has a vapor pressure of 7.6 bar at

21◦C. The boiling point of propane is -42◦C (-44◦F) which makes it vaporize as soon as

it is released from its pressurized container. Propane has an auto ignition temperature

of 480◦C. The propane used in our experiments was purchased from the company of

Air Liquid.

4.2.2 Si-Precursor: Silane

Silane is a colorless, pyrophoric gas. Being pyrophoric, leaks will immediately ignite

on contact with air. However, under certain conditions such as high humidity or rapid

release, it may not immediately ignite and may form pockets of gas, which may cause

a delayed vapor cloud explosion. This means it is self-igniting in air because it has

an autoignition temperature below ambient temperatures. Leaks of silane may cause a

fire or possibly form explosive mixtures in air. Some very small leaks will not give a

47

4.2 Gases CONTENTS

visible flame but may be detected by the presence of oxide powder buildup at the leak

point. The chief product of combustion is silicon dioxide SiO2. Silane fires must only

be extinguished by stopping the flow of silane to the fire. The lower flammable limit is

1.37 volume percent. Silane mixtures in the 3 to 4 percent range can exhibit pyrophoric

behavior. Its autoignition temperature is -50◦C. It has a density of 1.35 kg/m3 at 20◦C.

Its bpoiling temperature is -112◦C. Silane has a very high vapor pressure; at -3◦C, it

has a vapor pressure of 47 bar, which possibly develops reasonable concentrations at

high temperatures compared with chlorinated silanes like SiH2Cl2, SiHCl3 and SiHCl4,

which have lower decomposition temperatures and higher nucleation rates in hydrogen

dilutions as reported in [45]. The decomposition kinetics of silane is therefore very

important for systems with fast temperature rise, where the decomposition of silane do

not occur at low temperatures. Silane was purchased from Air Liquid company at two

volumetric concentrations of 1.5% and 5%.

4.2.3 Carrier Gas

In such open systems, convection heat transfer reveals through forced convection

and/or free convection. The velocity range in the susceptor plays a decisive role defin-

ing the convection mode. In forced convection, the internally imposed flow is generally

known, whereas in natural or free convection, the flow results from an interaction of

the density difference with their gravitational, or some other body force. The velocity

in natural convection is supposed to be slow compared with the velocity in the case

where forced convection dominates. In our case, this velocity of the process gases lies

in a range less than 0.0075 m/s, which is too slow for developing a forced convection

mode, which indicates a prevalence of free convection mode.

48

CONTENTS 4.2 Gases

In previous works that were presented in sec.2.3, H2, He and Ar were widely chosen

as carrier gases in SiC growth systems. In order to determine the optimal heat transfer

conditions in the hot-wall reactor, a convection based comparison was carried out for

all three gases individually using a susceptor geometry of H = 600 mm for its length

and a diameter of d = 130 mm. Firstly, the convection mode was investigated at as-

sumed uniform temperature of Twall = 2200◦C on the inner susceptor walls. The gases

properties were taken at an average temperature Tf = 1400◦C as presented in Table

4.1. The values of the table are showing that Ar has much lower thermal conductivity

compared with both H2 and He, where the dynamic viscosity and density of Ar is larger

than those of H2 and He. The inlet temperature Tin is assumed to be 600◦C, while the

gas is assumed to leave the susceptor at a temperature of T∞ = 2000◦C, which is 200◦C

below the wall temperature.

Gas Cp 103 ∗ ρ 106 ∗ η 103 * k

[J kg−1K−1] [kg m−3] [kg m−1s−1] [J m−1 s−1 K−1]

H2 15784 17.53 25.0 560.0

He 5196.6 34.8 55.0 460.0

Ar 520.7 347.9 65.0 52.6

Table 4.1: The carrier gases properties are given at 1400◦C,

where Cp is the heat capacity, ρ is the density, η is the dynamic

viscosity and k is the conductivity coefficient.

The convection mode can be distinguished by applying equation 4.1. If the condition

49

4.2 Gases CONTENTS

given by this equation is observed, free convection is assumed to be dominating.

GrHRe2H

>> 1 (4.1)

Where Reynolds number ReH is a dimensionless number that gives a measure of the

ratio of inertial forces to viscous forces and consequently quantifies the relative impor-

tance of these two types of forces for given flow conditions and Grashof’s numberGrH

which is a dimensionless number used for approximating the ratio of the buoyancy to

viscous force acting on a fluid. It frequently arises in the study of situations involving

natural convection. Both numbers are respectively given by equations 4.2 and 4.3.

Re =u ∗Dν

(4.2)

GrH =g ∗ β ∗ (Twall − T∞) ∗H3

ν ∗ k(4.3)

Where ν is the kinematic viscosity (calculated from η/ρ), β is 1/Tf , H is the reactor

length (600 mm), D is the hydraulic diameter given by 4*area/(π*d) for cylinders and

k is the thermal conductivity of the gases. The resulted values of the term Gr/Re2, are

much greater than 1 for all gases, as evaluated from the Gr and Re numbers listed in

Table 4.2, which matches the condition given in Eq.4.1 proving that free convection

dominates, meanwhile forced convection is accordingly assumed to be negligible. Now

the heat transfer can be calculated by equation 4.4 as found in [44].

QH =NuH ∗ k ∗∆Tin,wall

H(4.4)

Where ∆ Tin,wall is the difference between the inlet gas temperature (600◦C) and the

susceptor temperature (2200◦C) and NuH is Nusselt number of an upward flow in

50

CONTENTS 4.2 Gases

cylinders given by equation 4.5.

NuH = 0.52 ∗ (RaH)1/4 (4.5)

Where Ra stands for the Rayleigh number and is given by equation 4.6 [46]. In fluid

mechanics, the Rayleigh number for a fluid is a dimensionless number associated with

buoyancy driven flow (also known as free convection or natural convection). When the

Rayleigh number is below the critical value for that fluid, heat transfer is primarily in

the form of conduction; when it exceeds the critical value, heat transfer is primarily in

the form of convection. .

Ra =β ∗ g ∗ Cp ∗ ρ2 ∗H3 ∗ (Twall − T∞)

η ∗K(4.6)

The heat transfer coefficient (h) can be determined using equation 4.8. The values

calculated for h are given in Table 4.3 for all three gases.

h =k ∗NuD

(4.7)

Gas Re Ra*10−5 Gr*10−4 Nu

H2 0.68 10.5 0.38 16.64

He 0.62 7.5 0.42 15.31

Ar 5.24 558.0 30.8 44.94

Table 4.2: Re, Ra, Gr and Nu are unit-less numbers calculated

for H2, Ar and He.

By substituting the resulted values of the Re, Gr, Ra and Nu numbers listed in table.4.2

51

4.2 Gases CONTENTS

in the equations 4.4 and 4.7, the heat transfer and heat transfer coefficient can be calcu-

lated for free convection in all three gases as presented in Table 4.3. Upon those results,

it was found that H2 has the maximum heat transfer coefficient, which requests an en-

ergy of 24.85 kW to reach an outlet temperature of 2000◦C, while He and Ar require

less energy than H2 in order to reach the same temperature. Although the use of Ar

as carrier gas would lead to the lowest consumption rate of energy, but inhomogeneity

or thermal instability could be resulted due to its very low heat capacity. Therefore,

the higher thermal conductivity and heat capacity of He make it as a better alternative

of Ar and also H2, which requires higher consumption rate of energy than He. A heat

transfer rate of 19 kW was obtained using He. This value seems to be quite reasonable

in our case, taking into consideration that large amount of the energy will be dissipated

by the water and air cooled parts of the hot-wall reactor.

Gas h[W/m2*K] QH 10−3*[W/m2] 102*QH,carrier/QH,He %

H2 71.68 24.850 132.2

He 54.20 18.790 100

Ar 18.19 6.30 33.5

Table 4.3: The heat transfer rates are calculated in He, Ar and

H2 mediums. They were compared with each other taking the heat

transfer in He as datum. The maximum heat transfer coefficient

was resulted in Helium medium.

As previously explained, a flow rate of 6 slm was required for the process causing a

flow velocity of 0.0075 m/s. By setting such a velocity in the susceptor, laminar flow

and free convection heat transfer will be established. Higher flow rates might develop

52

CONTENTS 4.2 Gases

a forced convection heat transfer or combined free and forced convection mode. In

such a case, at a given instance of the flow stream there will be some delay until

the temperature and velocity have reached their maximum. The distances required

for an instance of the fluid to reach these two conditions are called: fully developed

temperature profile HT and fully developed velocity profile HV . HT and HV that are

given by the equations 4.8 and 4.9.

HT = 0.28 ∗D ∗ReD (4.8)

HV = 0.04 ∗D ∗ReD (4.9)

Re, HT and HV are calculated at three different helium flow rates of 6 slm, 24 slm and

42 slm, as presented in table 4.4. At the lowest flow rate of 6 slm (0.0075 m/s), the

distances HT and HV remain too small, where, as previously discussed, free convection

dominates. However, increasing the flow rate to 24 and 42 slm, results in increased

values of Re, HT and HV , which shows that the flow stream will require longer distance

to reach its maximum temperature and velocity in the reactor. Therefore, increasing the

flow velocity to relatively high values will result in a cooler flow stream, and therefore,

in lower growth temperatures. The flow stream temperature T(K ) for the gas center

can be calculated by equation 4.10, which can be applied in forced convection heat

transfer at high flow velocities where a value of 3.66 is used for the Nusselt number.

But, since increasing the velocity to a value of 0.0525 m/s results in a value of 683.8

for the term Gr/Re2, which is still much greater than 1, therefore, it could be resolved

that even at this velocity, the heat transfer mode remains dominated by free convection

53

4.3 Substrates CONTENTS

and the application of equation 4.10 is not valid at this flow condition.

V· (slm) Re HT (cm) HV (cm)

6 0.620 2.25 0.32

24 2.47 8.89 1.28

42 4.32 15.71 2.24

Table 4.4: Different flow rates of helium were applied. The cor-

responding distances of the fully developed temperature and ve-

locity are listed in this table. At a flow rate of 42 slm (velocity

of 0.0525 m/s), a distance of 15.79 cm was calculated for a fully

developed temperature profile, which indicates that increasing the

flow velocity will significantly influence the temperature profile.

(Twall − T (H)gas,center)

(Twall − Tgas,in)= exp (NuD ∗ k ∗ π ∗H/m. ∗ Cp) (4.10)

Where m. is given by equation 4.11.

m. = ρ ∗ π ∗ r2 ∗ u (4.11)

4.3 Substrates

As previously introduced in this chapter, seed-crystals are necessary for the epitaxial

growth of SiC using the HTCVD technique. Without a seed-crystal, amorphous or

polycrystalline films will be grown on arbitrary substrates. The polytype of the grown

SiC film can be determined by the polytype and the surface orientation of the seed-

crystal. For the seed-crystal fixation, the use of a suitable seed-holder is necessary. In

the hot-wall reactor design, it was planned to use a floating seed-holder, which can be

54

CONTENTS 4.3 Substrates

hanged horizontally inside the susceptor. These points are discussed in the following

part of this chapter.

4.3.1 Seed-holder

In non-seeded growth experiments, SiC was directly deposited on substrates made of

graphite foil, which were formed in different shapes as seen in Figure 4.3. In seeded

growth experiments, the substrate holder was made of hard graphite with a density of

1.83 g/cm3. Different versions of seed-holder were applied in the growth experiments

and improved corresponding to the growth results. More details about the applied

seed-holders in this work are presented in section 5.2.4.

Figure 4.3: The firstly used seed-holders

were cut out of a graphite foil. a) Vertically

hanged holders of long stripes. b) Hori-

zontally hanged holders with disc shape.

4.3.2 Substrate Polytype

On-axis, 2 inch 6H-SiC wafers with 250 µm thickness were purchased from SGL Car-

bon and divided manually into six pieces with triangular shapes for their initial appli-

cation in the epitaxial growth experiments, where epitaxial growth of SiC is expected

to result in its cubic polytype, as explained in section 2.3.3. For the homoepitaxial

growth of SiC, small 3.5◦ and 8◦ off-axis surfaces crystals were purchased with the

dimensions of 6×6×0.25 mm and applied for targetting the homoepitaxial growth of

55

4.4 Deposition Mode CONTENTS

6H-SiC.

4.3.3 Seed Adhering

The use of a free floating seed-holder, where the gas flow is in upward direction, re-

quires fixing the seed-crystal on the lower side of the seed-holder. The small size of the

seed-crystals and the required purity of the process make it difficult to mechanically

fix the seed-crystal, especially by the use of additional parts or materials. This problem

was previously solved by Kordina in [31], who fastened the seed-crystal on the seed-

holder using molten glycol. This method was successfully followed in our laboratory

using molten sugar. A thin layer of sugar was put between the seed-crystal and the

seed-holder, which was heated for around 10 minutes at a temperature of 250◦C until

the sugar was completely decomposed.

4.4 Deposition Mode

4.4.1 Temperature Profile

The hot wall reactor was designed for a maximum temperature of 2500 K. The temper-

ature was controlled by the output of the MF-generator, which is capable to produce an

output power up to 50 kW. The temperature on the susceptor walls is not constant on

its whole length due to the geometry of the inductive coil and the high rate of heat dis-

sipation on both ends of the susceptor. The optical access designed at the bottom of the

reactor enabled the temperature measurement along most of the susceptor inner space.

A mirror made of Aluminum was manufactured and polished to be used for reflecting

the rays emitted by the hot object (inner susceptor walls, seed-holder or seed-crystal)

to the pyrometer. This mirror is fixed on a bench, which enables flexible adjustment of

56

CONTENTS 4.4 Deposition Mode

Figure 4.4: A half view of the reactor cross section is represented on the left side, while on

the right side, the temperature distribution profile is illustrated in relation with the position on

the susceptor walls.

57

4.4 Deposition Mode CONTENTS

the mirror. The application of such a design allowed the monitoring of the temperature

along most area of the susceptor inner surface, which can be considered as a novel

improvement of the temperature measurement in HTCVD systems, where the temper-

ature was usually measured on small spots of the outer cylinder of the susceptor or on

the upper side of the lid, where the seed-crystal is fastened on its bottom.

Figure 4.5: Left: Picture taken through the optical access and shows the seed-holder (made

of graphite foil) at a temperature of 2000◦C. Right: Temperature profile of the seed-holder

starting from the center point and ending at the maximum radius of the substrate.

The temperature profile of the susceptor could be measured at a pressure of 300 mbar

under two different flow conditions, firstly without flow, and secondly under a flow

velocity of 0.0075 m/s at a Helium flow rate of 6 slm. The measured temperatures are

illustrated versus their individual z-positions (60 cm length) as seen in Figure 4.4. The

temperature profile has a non-constant distribution with a maximum wall temperature

of 2180◦C. This temperature was recorded for a point, which is coincident with the

middle length of the inductive coil, which indicates that the vertical position of the

inductive coil (in regard with the susceptor position) has a significant influence on the

temperature profile. The helium flow was found to have a negligible effect on the tem-

58

CONTENTS 4.4 Deposition Mode

perature distribution, which indicates that the applied flow velocity is still too slow for

decreasing the heat transfer coefficient and consequentially changing the heat trans-

fer mode from free to forced convection. The temperature measured at a z-position

of 5 cm below the outlet flange was 1700◦C, whereas a temperature of 1800◦C was

measured at a z-position of 46 cm below the outlet flange.

During heating, the seed-holder temperature was continuously measured by the pyrom-

eter till the desired growth temperature is reached. Once the desired value is achieved

and became stable, the precursors were then fed into the reactor. The silane SiH4 is

known for its pyrolysis at temperatures close to 1000◦C. A temperature range of 800-

1200◦C is usually reached in the duct causing an early or preliminary pyrolysis of SiH4

causing a dusty sight in the inner space of the susceptor, through it a continuous tem-

perature measurement is not possible, therefore, the SiH4 flow must be shortly stopped

during the deposition until the dust disappears from the reactor space, thereafter the

temperature measurement can be carried out. The temperature distribution was thus

measured on a seed-holder, which is hanged 18 cm below the outlet flange, as illus-

trated in Figure 4.5. The pyrometer has recorded a maximum temperature of 2025◦C

on the center of the seed-holder, while a temperature difference of around 50◦C be-

tween the center point and the substrate holder perimeter was recorded.

4.4.2 Geometrical Setup

In such a vertical flow hot-wall reactor, where the carrier gas and the precursors are in-

jected through the duct in an upward direction towards the stagnant surface of the sub-

strate, suitable positioning of the substrate holder is required. In non-seeded growth,

long stripes made of graphite foil were initially used to indicate the vertical range

59

4.5 Experimental Procedure CONTENTS

Figure 4.6: Schematic description of different substrate positioning. a) long stripe is hanged

vertically. b) long stripe is hanged with a small tilt angle to the susceptor centerline. c)

Circular substrate hanged horizontally stagnating the flow.

whereon polycrystalline SiC will deposit. This range was used as an indicative for the

initial substrate position in epitaxial growth experiments. Additionally, other graphite

stripes were hanged with a vertically inclined plane in order to cover a wide range of

the susceptor space. Figure 4.6 illustrates three different flow geometries.

In epitaxial growth experiments, the stagnation flow geometry was further changed

by adding an aperture below the substrate, whereby the flow velocity is increased and

consequentially, more of the flow stream can be focused on the stagnant surface of the

seed-holder. Different aperture geometries are illustrated in Figure 4.7.

4.5 Experimental Procedure

As previously mentioned, in seeded growth experiments, the seed-crystals were fixed

on graphite seed-holders. The seed-holder is conducted into the HWR by means of

graphite cords. Thereafter, the reactor was tightly sealed. The explosive nature of

SiH4 in air requires tight sealing of the reactor assembly to ensure safe deposition pro-

60

CONTENTS 4.5 Experimental Procedure

Figure 4.7: Three different stagnation flow geometries. a) The substrate is hanged horizon-

tally. b) An aperture is fixed a few centimeters below the substrate. c) A conical aperture is

used to avoid flow turbulences.

cess. In order to detect any leakage before starting the experiments, the reactor was

pumped out to a low pressure of 10 mbar, which was kept constant at this value for

testing leakage rate. Leak detection is a control method used to identify, monitor, and

measure the unintentional entry or escape of fluids and gases, usually from pressurized

systems or into empty enclosures. Leaks can move from the inside of a component or

machine into the outside, or penetrate from the outside in, due to differences in pres-

sure between two regions. Most leak detectors are primarily responsible for locating

the leak, determining the extent or rate of leakage, and keeping track of increases or

decreases in leakage. Leak detection is highly important in industrial systems that rely

on sensitive components or equipment with the potential for being damaged by exter-

nal contaminants. Leak testing and detection are implemented to prevent material or

energy loss, improve a manufacturing systems reliability, and reduce the risk of envi-

ronmental contamination. In our case, a helium leak detector was used to detect the

parts or the portions that are responsible for the leak. Since, O-rings were used to seal

61

4.5 Experimental Procedure CONTENTS

some parts in the reactor, so a leak rate of 10−9 mbar * l/s was expected. This step

is followed by a heating period of around 30 minutes. Once the growth temperature

required is reached on the seed-holder, the deposition can be started by feeding the re-

actor with the gas mixture. Once the deposition period is finished, all flow valves must

be closed(except those used for cooling water) and the MF-generator is switched off.

After a cooling period of 60 minutes, the seed-holder can be taken out of the reactor

and the seed-crystal can then be separated from the seed-holder for its investigation.

62

Chapter 5

Deposition of SiC

5.1 Non-seeded Growth of SiC

Non-seeded growth of SiC can be achieved using the same concept like the one ap-

plied in the Lely method, which is presented in section 2.2.2. This method was used

to observe the growth of polycrystalline SiC at high temperatures and find optimum

conditions including the substrate position. Long stripes of graphite foil were hanged

vertically in the reactor whereon, SiC is expected to be grow. The deposition of poly-

crystalline SiC was also studied on circular substrates made of graphite foil, which

were hanged normally to the upward flow direction.

5.1.1 Observation of SiC Growth on Graphite Stripes

In order to observe the growth of SiC using our HTCVD system, long stripes made

of graphite foil were hanged vertically in the reactor with slight inclination in order to

extend the area whereon, SiC can be possibly deposited. The stripe has a length of 60

cm and a width of 5 cm. A maximum temperature of 2160◦C was measured on the

63

5.1 Non-seeded Growth of SiC CONTENTS

susceptor walls, where the pressure was set to a value of 300 mbar. A flow rate of 4

slm of diluted SiH4 in helium with 1.5% (60 sccm) and 30 sccm C3H8 have been fed

into the reactor for 30 minutes. The total gas flow rate was set to 5 slm. 1 slm of this

flow is used for cooling the glass disc of the optical access.

Figure 5.1: Long stripe of graphite foil

hanged vertically and deposited with SiC.

A film of smooth, black powder was rec-

ognized on the lower area of the substrate,

while a brilliant surface could be noticed

on the upper area of the stripe.

A film grown on the graphite foil was noticed on the top area of Figure 5.1, which is

considered as the first evidence for the growth of SiC by the present HTCVD system.

However, the deposition range of SiC was indicated within a vertical length of 5-10 cm

starting at a position of 15 cm below the outlet flange. Unlike this area, where a bright

green film is visible, a dark area was noticed on the bottom of the graphite foil. On this

area, no optical evidence that proves the growth of SiC crystallites could be observed,

which indicates improper growth conditions in this range.

The top area was magnified by optical microscopy as seen in Figure 5.2, where random

growth of small crystallites in form of bars is visible in this image. Such non-epitaxial

(polycrystalline) film growth can be described as a polycrystalline SiC.

64

CONTENTS 5.1 Non-seeded Growth of SiC

Figure 5.2: Image taken with an optical

microscope. Small crystallites in form of

small bars were grown on the graphite

foil. No preferred orientation of the grown

crystallites is visible.

Figure 5.3: X-ray diffractogram of the grown polycrystalline film on the graphite foil. The

X-ray diffractograms of graphite, silicon and 6H-SiC are also plotted. The reflections 1,2,3,4

and 5 of the sample indicate the growth of the cubic SiC polytype. No reflections could be

observed at 34.2◦ or 38.2◦, which observes that 6H polytype was not included in the grown

film.

65


The film was also analyzed by XRD as seen in Figure 5.3. The positions of the Bragg

reflections were compared with powder diffraction patterns of of different SiC poly-

types. The most important peaks of the sample were assigned the numbers 1,2,3,4

and 5 and compared with the diffraction pattern of the diffractogram discussed by Kli-

menkov in [47]. The most intense peak 1 at θ = 35.6◦ of 6H-SiC (102) is nearly at

the same position as the strongest 3C-SiC peak (111), therefore it cannot be used for

identification. Thus, the reflections 2,3,4 and 5 that are observed respectively at θ =

41.2◦, 61◦ of 3C<220>, 72◦ of 3C<311> and 77◦ are overlapping in both 3C and 6H

patterns, but since no reflections could be observed at θ = 34.2◦ or 38.2◦, which are

necessary for the identification of the 6H polytype, it is likely that the grown film is of

the 3C-SiC polytype.

5.1.2 Stagnation Flow Geometry

The observation of polycrystalline SiC growth, was followed by finding the vertical

level at which SiC intensively and optimally grows. In the previous experiment it

was noticed that SiC tends to deposit within a vertical range of 5 to 10 cm length on

the graphite stripe. This range starts at a distance of 15 cm below the outlet flange,

therefore the substrate was placed at this level (15 cm below the outlet flange). The

applied (stagnation flow) geometry was previously introduced in Figure 4.6-c.

Figure 5.4: Homogeneous film growth of

brilliant, green crystallites of polycrys-

talline SiC.

66

CONTENTS 5.1 Non-seeded Growth of SiC

Figure 5.4 shows a substrate made of graphite foil with 10 cm diameter. SiC was

deposited on it for 30 minutes at a pressure of 300 mbar and a maximum reactor tem-

perature of 2160◦C. Both concentrations of SiH4 and C3H8 were set to both values of

0.484 and 0.121 mol/m3 at flow rates of 60 sccm, 15 sccm, respectively, while the total

flow rate was set to 5 slm with a flow velocity of 0.0063 m/s. The deposition of a

green brilliant film of polycrystalline SiC was optically observed on the substrate. The

homogeneity of the grown film is an indicative for reasonable growth conditions and

precursor concentrations.

5.1.3 Fastening of the Seed-crystal

Fixing the seed in the HTCVD reactor is one of the problems, which had to be over-

come in the early phase of this work. The substrates or the seed-crystals were fastened

on the seed-holder using the method introduced in section 3.3.3. 6H-SiC wafers are

expensive, therefore only small pieces of the wafers were used in the experiments to

reduce the experimental cost. As seen by the picture of Figure 5.5, the seed-crystal

was fastened to a seed-holder made of graphite foil with a circular shape of 9.5 cm

diameter, which was hanged at a distance of 18 cm below the outlet flange. A maxi-

mum temperature of 2180◦C was recorded on the reactor walls, where a temperature

of 1850◦C was measured on the seed-holder. Both concentrations of SiH4 and C3H8

were set to 0.573 and 0.229 mol/m3 at flow rates of 100 sccm, 40 sccm, respectively,

while the total flow rate (including the carrier gas) was set to 7 slm leading to a flow

velocity of 0.009 m/s at a pressure of 300 mbar.

Here, only the film grown on the non-seeded area is discussed, it was analyzed by

means of SEM. A small crystallite of around 6 µm width was magnified by the SEM

67


Figure 5.5: Left, picture shows an SiC seed-crystal, which was fixed on circular graphite foil.

Right, SEM image shows the surface morphology of the film grown on the non-seeded area of

the seed-holder.

image seen in Figure.5.5-b. Grooves with parallel traces could be noticed on the ir-

regular surfaces of this crystallite, which indicates an oriented growth of the crystallite

epilayers. Both elements Si and C were detected by EDX analysis as seen in the spec-

trum shown in Figure 5.6. A composition ratio of (C/Si: 56/44) was evaluated on the

film, where an error of ±0.1-0.5% can be considered for the values resulted by the

EDX evaluations. This ratio differs slightly with the ratio obtained for an unused SiC

wafer (C/Si: 48/52), which observes a slightly excessive carbon composition, which

seems to be a result of high C3H8 concentration.

The picture shown in Figure 5.7 is taken from the seed-holder backside after 30 min-

utes of deposition, whereon yellow powder was accumulated during this period. This

powder was analyzed by XRD as shown in Figure 5.8. It shows that all the reflexes

obtained by the powder sample, at θ = 41.2◦, 61◦ of 3C<220>, 72◦ of 3C<311>

and 77◦, are coincident with the reflexes of cubic SiC pattern, which proves that this

powder if of the 3C-SiC polytype. Thus, comparing the positions of the resulting re-

68

CONTENTS 5.2 Seeded Growth of SiC

Figure 5.6: EDX-spectrum observes a film composition of both carbon and silicon elements.

flections by the sample X-ray diffractogram with the positions of the reflections of the

6H pattern, no further reflections could be observed at θ = 34.2◦, 41.2◦ or even at 65◦,

which indicates that the 6H polytype was not grown.

Figure 5.7: The picture is taken from the

substrate holder backside and shows a

free, undesired growth of yellow SiC pow-

der, which was accumulated in form of

small bars.

5.2 Seeded Growth of SiC

The results discussed in the previous part of this chapter have discussed the film growth

on non-seeded substrates, which resulted in the growth of polycrystalline SiC. For the

epitaxial growth of SiC, the use of a seed-crystal is necessary when single crystalline

69

5.2 Seeded Growth of SiC CONTENTS

Figure 5.8: The X-ray diffractogram of the powder sample was compared with the powder

diffraction patterns of 3C and 6H-SiC. All the peaks of the sample pattern are coincident with

the peaks has cubic 3C-SiC.

70


growth is desired using the HTCVD technique. The results of the films grown on

seed-crystals with different surface orientations are discussed.

5.2.1 Substrate Surface Treatment

Graphitization of the seed surface in the HTCVD technique is one of the known, severe

problems, which is usually generated during the initial stages, especially during heat-

ing. Graphitization is caused due to the deposition of the carbon decomposed from the

susceptor into the gas-phase, which later deposits on the seed-surface as by the decom-

position of the substrate due to incongurrent evaporation. The quality of the epitaxial

layer depends not only on the CVD growth conditions, but also on the substrate quality.

The surface of commercially available substrates contain a variety of defects that will

disturb the growth of the epitaxial layer reducing its quality. Such defects that orig-

inate from the substrate can not be eliminated before growth start, but good surface

treatment may inhibit further defects that can be caused by carbon transport from the

graphite parts to the seed surface. Surface preparation of SiC wafers was frequently

carried out using hydrogen [48] or hydrogen/propane [49]. In this work, different flow

procedures were applied during heating in order to reduce the transport of the carbon

decomposed from the susceptor material to the seed surface. Those procedures were

studied upon the SEM and EDX analysis of the surfaces of the heated seeds. In all

procedures the substrates were heated for 30 minutes till a temperature of 1900◦C was

reached. Firstly, the substrate was heated under a pressure of 800 mbar and a helium

flow of 6 slm. Secondly, only a treatment in a hydrogen flow of 200 sccm was applied

at a pressure of 800 mbar. Finally, the reactor was just pumped out till a pressure less

than 10 mbar was reached, thereafter heating was started without any flow. The surface

morphology of the three seeds are shown in Figure 5.9. The SEM images of the three

71


seed surfaces could observe the growth of small deposits with random distribution.

Figure 5.9: Three SEM images taken for 3 different seed crystals that were heated up to the

actual growth temperature in a period 30 minutes. Image a) shows a seed surface morphology

after heating in helium atmosphere, which was continuously supplied with 6 slm at 300 mbar.

Image b) shows seed surface heated under a hydrogen flow rate of 200 sccm at 300 mbar.

Image c) shows the surface morphology of heated seed at low, constant pressure of 10 mbar

without any addition of gases.

The lowest carbon concentration (C/Si ratio) was evaluated on the seed surface heated

using the third procedure. A composition ratio of C/Si:72/28 was evaluated for the

seed surface using this procedure. The application of both other heating procedures,

resulted in a carbon-rich seed surface. A composition ratio of C/Si:80/20 was evaluated

for the seed heated due to the first procedure, while a ratio of C/Si:88/12 was evaluated

by applying the second procedure. According to this result, it can be concluded that,

firstly, heating the reactor at low pressure and zero flow rate reduces the transport

of graphite from the susceptor to seed surface. Secondly, the carbon contamination

of the surface can not be completely inhibited using any of those flow procedures,

which requires further improvement of the flow procedure or applying further etching

methods.

72


5.2.2 Indication of Optimum Growth Temperature

The non-constant temperature distribution recognized by the temperature profile pre-

sented in section 3.4.1, indicates a significant relation between the seed-holder po-

sition and its corresponding temperature. In other words, it could be resolved that

the seed-holder temperature (growth temperature) is a function of distance (H) mea-

sured between its position and the outlet flange. Therefore, seeded growth experiments

were carried out at different seed-holder positions in order to increase or to decrease

the growth temperature. In this section, seeded growth results are discussed for films

grown at three different seed-holder positions. Table 5.1 lists three distances of H, their

corresponding temperatures and the EDX composition ratios. The concentrations of

the SiH4 and the C3H8 were set to 0.667 and 0.200 mol/m3 at flow rates of 100 sccm

and 30 sccm, respectively. The helium flow was set to 6 slm resulting in a flow veloc-

ity of 0.0075 m/min at a pressure of 300 mbar. The deposition was carried out for 20

minutes.

Experiment H (cm) Tsubstrate (◦C) Composition Ratio (C/Si)

a) 14 1850 47/53

b) 16 1950 47/53

c) 18 2035 -

Table 5.1: Table listing three different substrate positions and the corresponding temperatures

and composition ratios of the film’s surfaces.

The surface morphologies of the resulting films were firstly compared with each other

based on optical microscopy, see Figure 5.10. A film of small crystallites was grown

73


on the seed-crystal of experiment (a). The increased temperature caused by shifting

the seed-holder 2 cm towards the hot zone in experiment (b), resulted in the growth

of crystallites with hexagonal boundaries on a large area of the substrate. The same

surface morphology found on the film obtained in (a), was also recognized inside the

boundaries of the hexagonal areas obtained by experiment (b). The hexagonal shapes

obtained in (b) denote a relative enhancement of the film morphology towards epitaxy,

which is a result of the increased growth temperature. Further increase of the tem-

perature due to lower positioning of the seed-holder, as in experiment (c), resulted in

the growth of large crystallites with plain surfaces. The morphology resulted in (a)

and (b), where small crystallites were compactly grown, was completely disappeared

in the film obtained in (c). Upon those results, it could be concluded that a growth

temperature of approximately than 1950◦C is required to advance the growth condi-

tions. Indeed, lower positioning of the substrate towards the hot zone results in higher

temperature of the seed-holder, which consequentially leads to improving the epitaxial

growth conditions.

Figure 5.10: Images taken with optical microscope and show the surface morphology of three

films grown at different growth conditions. Those conditions are presented in table 5.1 and

explained in the text.

74


For further investigation of the surface morphology and the composition of the grown

films, SEM and EDX analysis were carried out for both films obtained by experiments

(a) and (b). A composition ratio of (C/Si: 47/53) was evaluated for the film obtained

in experiment (a), which indicates a reasonable precursor concentrations. The growth

of polycrystalline SiC could be observed by analyzing the surface morphology of both

SEM images shown in Figure 5.11.

The film obtained in experiment (b) was thus analyzed by means of SEM. A small area

of 250 µm width, was magnified to clearly show its morphology details, see Figure

5.11-b1. One hexagonal structure is clearly seen in the SEM image of Figure 5.11-

b2, which has a width of 8-9 µm. The appearance of such a new structure within

the surface morphology could be considered as an indication for a relative advance

of the growth conditions, whatever that the resulting film is optically recognized as

a polycrystalline film, as seen by the image of Figure 5.12-b3. This image observes

the growth of compact film composed of small SiC crystallites. The EDX analysis

evaluated a composition ratio of (C/Si:47/53) for this film, which is approximately

equal to the composition ratio evaluated for an unused 6H-SiC seed-crystal of C/Si:

48/52. This result indicates an appropriate gas-phase composition that is established at

the applied precursor concentrations. Generally, it could be concluded that increasing

the temperature to a value of 1950◦C, leads to enhancing the growth conditions that

are necessary for achieving an epitaxial growth mechanism.

5.2.3 Improved Flow Geometry

In the previous experiments, it was noticed that large amount of SiH4 decomposes in

the low temperature zone, which leads to particle building that later fall on the glass

75


Figure 5.11: Two SEM images of the film grown in experiment (a). Image (b) is a magnified

area of image (a). The film is composed of compactly grown crystallites.

Figure 5.12: SEM images of the film obtained in experiment (b). Image (a) shows the sub-

strate surface, where hexagonal structures are visible. (b) Magnification of one hexagonal

structure. (c) Shows the internal morphology of the hexagonal structure shown in image (b).

Compact growth of small, individual crystallites are observed by this image.

76


plate of the optical access. The installation of an aperture below the seed-holder, as

presented in section 3.4.2, was suggested as an alternative flow geometry that can be

used to force the most amount of the flow stream to be stagnated by the seed-holder.

The use of the aperture is involved in this section only. An aperture made of graphite

foil, which has an inner diameter of 50 mm and an outer diameter, which is equal to the

susceptor inner diameter of 130 mm, was installed inside the susceptor at a distance

of 28 cm below the outlet flange. The seed-crystal was fixed on a seed-holder made

of graphite foil and placed 24 cm below the outlet flange. A temperature of 1800◦C

was measured on the seed-holder, where the aperture temperature exceeded 2100◦C.

The concentration of both precursors of SiH4 and C3H8 were set to 0.573 and 0.229

mol/m3 at flow rates of 100 sccm and 40 sccm, respectively. A total flow rate including

the helium carrier gas was set to 7 slm leading to a flow velocity of 0.6 m/min. The

deposition was carried out for 20 minutes at a pressure of 300 mbar.

Figure 5.13: Triangular shaped crystal-

lites were grown on an on-axis seed sur-

face. The edges of the crystallites are

shown to be grown towards three certain

directions, which reveals the influence of

the seed surface orientation on the growth

mechanism.

The surface morphology of the grown film is shown by the SEM image of Figure

5.13. Several crystallites with triangular shape were grown on an on-axis 6H-SiC

seed. The crystallite edges were grown into three parallel directions. Such a mor-

77


phology is of great importance, since it showsthe possibility of growing SiC epilayers

by our HTCVD system. The EDX-analysis evaluated a composition ratio of (C/Si:

51/49), which observes a slight excess of carbon concentration comparing it with the

composition of C/Si: 48/52, which was evaluated for a non-used 6H-SiC wafer. This

indicates a relatively high concentration of the C3H8 added to the process.

5.2.4 Seed-holder Improvement

Figure 5.14: Four different geometries of the seed-holder (SH). a) SH has a diameter of 10

cm, b) Hollow SH with 5 cm diameter and 25 mm hight. The graphite cord is normally tied

through both holes shown. c) SH with 5 cm diameter and 25 mm hight. d) Improved version of

SH with 5 cm diameter and 12 mm hight. This SH can be hanged via a hole drilled through its

ridge.

Graphite foils did not show a promising performance using it as a material for the seed-

holder, since the seed-crystals were strongly bonded to the holder surface or damaged

at their adhered side, which usually resulted in some mass reduction of the seed-crystal.

Hard, dense graphite was used instead of graphite foil in order to solve this problem.

78


Four geometries to the seed-holder design were studied and gradually improved upon

the growth results. The four seed-holder types a,b,c and d are shown in Figure 5.14.

Firstly, applying seed-holder a), which has a relatively large diameter of 10 cm led to

hardly bonded seed and holder, which consequentially caused the damage of the seed-

crystal while detaching it. Additionally, only the central area of the seed-holder was

coated, meanwhile the outer area near the perimeter remained free of deposition. This

problem was later eliminated by the use of smaller seed-holders with a diameter of

50 mm, which indicates that the seed-holder size seems to have a significant effect on

the diffusion layer near the substrate surface. However, in order to solve the bonding

problem, the seed-holder geometry was improved in such a way, where the contact

area between the seed and the holder is minimized. Accordingly, the seed-holder b)

was made in beaker shape, as seen in Figure 5.14-b. In this case, a complete wafer with

a diameter of 50 mm must be adhered on its cylindrical walls, which have a thickness

of 2 mm. Such a type of seed-holder was applied in one experiment, where a 2-inch

diameter SiC wafer with on-oriented surface was deposited for 2 hours. The seed-

holder was hanged by means of a graphite cord 18 cm below the outlet flange and

heated till a growth temperature of 1900◦C was reached. The precursor concentrations

of SiH4 and C3H8 were set to 0.5736 and 0.172 mol/m3 at a flow rates of 100 sccm and

30 sccm, respectively. The total flow including the carrier flow rate was set to 7 slm

leading to a flow velocity of 0.0088 m/s.

A rough, green film was found in the middle area of the substrate shown in Figure

5.15, while a smooth, tapered area was noticed near the perimeter, which observes the

decomposition of the seed material in this region. However, the growth of a polycrys-

talline film could be observed by means of the SEM image shown in Figure 5.16-a.

79


Figure 5.15: 2-inch wafer fastened on

the hollow seed-holder. The circumferen-

tial area looks flat, while a rough growth

morphology could be noticed in the mid-

dle of the wafer after deposition.

The grown crystallites are shown to be oriented towards three certain directions, which

resembles a remarkable influence of the surface orientation of the seed polytype (6H,

on-axis) on the growth mechanism. The seed-crystal was broken into relatively large

pieces while detaching it from the substrate holder. The cross section area of one of the

broken pieces is shown in the SEM image seen in Figure 5.16-b. A trace distinguishing

the film from the original seed-surface was recognized on this image, which was used

for measuring the film thickness. By measuring the distance between this trace and

the film surface, a relatively thick film of 40 µm could be observed corresponding to a

growth rate of 20 µm/h. Although this seed-holder type simplified the detachment of

the seed-crystal in order to measure the film thickness for the growth rate determina-

tion, the damage of the substrate on its backside could not be inhibited as observed on

the SEM image shown in Figure 5.16-b.

A further disadvantage of the hollow seed-holder type was noticed during the tempera-

ture measurement, since the pyrometer has recorded instable values for the temperature

during heating. This problem could be suppressed using another seed-holder of type

(c), which has a simple shape of a can, diameter of 50 mm and a thickness of 25 mm.

This type of seed-holder was further optimized to the type (d), which is approximately

50% lighter than the (c) type providing an attractive weight that enables its connection

80


Figure 5.16: a) SEM image taken of a small portion of the middle area of the seed surface,

where the growth of small longitudinal crystallites in three growth directions is visible. b)

SEM image showing the side view of the cross section area of the deposited seed-crystal. A

film thickness of 40 µm could be observed on this image. The adhered side of the wafer looks

damaged.

with the magnetic suspension balance, which has a maximum measuring capacity of

80 g.

5.2.5 Improvement of Growth Conditions for Epitaxy

Figure 5.17: On-Axis SiC substrate fixed

on a graphite seed-holder of 50 mm diame-

ter. A plain, transparent surface morphol-

ogy is visible. Growth defects can be rec-

ognized within a small area on the right of

the seed surface.

The results introduced in section 4.2.2 have shown a significant enhancement of the

81


film quality due to shifting the seed-holder position towards the hot zone and conse-

quentially increasing its temperature. This result resolves the importance of the growth

temperatures applied in relation with the vertical position of the seed-holder. Accord-

ingly, the seed-crystal was fastened on a graphite seed holder, as seen in Figure 5.17,

and hanged at a distance of 30 cm below the outlet flange. The seed-crystal has a tri-

angular surface and is a part of an on-axis 6H-SiC wafer. A temperature of 1995◦C

was measured on the seed-holder at this position. A maximum temperature of 2050◦C

was recorded at a point, which is 35 cm below the outlet flange. The SiH4 concen-

tration was set to 0.996 mol/m3 at a flow rate of 150 sccm, while the concentration

of C3H8´was set to a relatively low concentration of 0.066 mol/m3 at a flow rate of

10 sccm, because a very low ratio of (C3H8/SiH4:1/8) was established by O.Kordina

[22], who recommended to start with such low concentrations of C3H8. The carrier gas

flow rate was set to 6 slm providing a total flow velocity of 0.0075 m/s. The epitaxial

growth of SiC by HTCVD was achieved by Elisson in [32] at high growth rates of 1

mm/h at an established pressure, which is close to atmospheric pressure. Therefore,

the pressure was also set to a value of 800 mbar, which is relatively lower than the at-

mospheric pressure in order to avoid sudden pressure increase above the atmospheric

pressure, which might lead to the explosion of the quartz tube. The deposition was

then carried out for two hours.

By the optical inspection of the seed-crystal, the appearance of the unused seed-crystal

remained partially unchanged. The seed surface has kept its original green color and

transparency. However, on the right side of the seed surface, a small portion with

rough morphology was recognized the image a1 shown in Figure 5.18. An area of

1 mm2 is magnified by the SEM image shown in Figure 5.18-a1. Growth defects

82


Figure 5.18: SEM images of two films grown at different propane concentrations. Image (a1)

shows growth defects grown in an area of 1 mm. The defects are magnified in image (a2),

which observes an oriented defect growth. Image b) shows a plain surface morphology of film

deposited at higher propane concentration.

are visible on a large area of the film, which seems to be caused due to improper

precursor concentrations. An oriented defect growth could be observed by further

magnification of the surface morphology, as seen by the image a2 of Figure 5.18, where

the edges of the defect areas are seen to be orientated towards three certain directions.

Such an oriented morphology of the grown defects reflects the effect of the original

seed polytype on the grown film structure, which indicates that the process is pushed

towards epitaxy. In fact, different reasons can be responsible for such a growth defect.

Seed surface contamination with carbon during heating, unsymmetrical temperature

gradients on the seed crystal or even the low concentration of C3H8 are suspected to

be the reason that is responsible for obtaining such a result. Therefore, the surface

morphology and the growth rate was further investigated by SEM and EDX at higher

concentrations of C3H8.

The C3H8 concentration was increased to a value of 0.099 mol/m3 (15 sccm), which

resulted in a remarkable improvement of the surface morphology as seen on the SEM

83


Figure 5.19: SEM image of a vertical cut

in the substrate. The total substrate thick-

ness after deposition is 303µm, which cor-

responds to a growth rate of 26µm/h.

image shown in Figure 5.18-b. This shows a defect free, plain surface morphology. The

growth rate was initially evaluated by subtracting the seed weights after and before the

growth experiment, which was thus validated by means of the optical microscopy or

SEM analysis. A growth rate of 17 µm/h (see table 5.2) was achieved in the first exper-

iment, where low concentration of C3H8 (0.066 mol/m3) was used, while in the second

experiment, a growth rate of 24 µm/h resulted by increasing the C3H8 concentration to

0.099 mol/m3. The latter growth rate was evaluated by the weight subtraction, while

a total seed thickness of 303 µm could be measured using the SEM image of Figure

5.19, which makes a growth rate of 26.5 µm/h. A deviation of 7% was calculated for

both evaluated growth rate values in the second experiment. Indeed, increasing the

C3H8 concentration in the second experiment enhanced the growth rate by 56%. Thus,

the growth defects observed at very low C3H8 concentration of 0.066 mol/m3, disap-

peared of the film obtained by the second experiment (see Figure 5.18-b), where the

concentration of C3H8 was slightly increased by 50% to 0.099 mol/m3.

The crystallography of the film grown by the second experiment was further analyzed

by means of Low Energy Electron Diffraction (LEED) at the Institute of Experimental

Physics. Figure 5.20 shows two known patterns of hexagonal SiC with (3×3) and

(2×2) reconstructions. Despite of the unclear or bad appearance of atom spots, which

84


is mostly caused due to non-planer film surface, both patterns are observed for both 3C

and 6H polytypes of crystalline SiC as reported in [50] and [51]. This result resolves

the success of the epitaxial growth of crystalline SiC.

Figure 5.20: Left: LEED pattern of (3×3)obtained at 564 eV. Right: LEED pattern of (2×2)

obtained at 364 eV.

The film prepared by the second experiment was further analyzed by means of XRD,

see Figure 5.21. The substrate sides were covered by an aluminum foil to exclude the

diffraction of the entire layers of the original substrate depth within the resulting film.

Since these layers might belong to the original layers of the seed-crystal and not to the

grown layers. The reflections of the aluminum were observed at θ = 38◦, 45◦ and 65◦.

The X-ray diffractogram obtained for the film surface has shown that sharp peaks were

resulted at the reflection angles 37◦ and 76◦ with high intensities. Comparing this result

with the diffractogram obtained by the XRD analysis of an unused on-axis 6H-SiC

seed-crystal, it can be found that all its reflections are also obtained by the film pattern.

Only the reflex observed at θ = 42◦, indicates the growth of 3C polytype. Although

85


Figure 5.21: Two X-ray diffractograms of (a): unused 6H-SiC seed-crystal with on-axis sur-

face, and (b) Deposited seed-crystal on on-axis seed surface. The seed edge was covered at its

side with Al (its reflection at 38◦). Further reflection at 42.2◦ identifies the growth of 3C-SiC.

The reflection observed at 34.2◦ indicates the growth of 6H polytype.

86


this reflex is supposed to be observed by both polytypes 3C and 6H as reported in [47],

it is here more expected to stand for a 3C reflection, since such a reflection was not

observed in the X-ray diffractogram of the original 6H-SiC seed-crystal. This result

meets the known theory of the heteroepitaxial growth of 3C polytype on on-oriented

surfaces of SiC.

Sample C3H8 (mol/m3) Growth Rate (µm/h) Composition Ratio (C/Si)

Unused seed - - 48/52

a 0.066 17 48/52

b 0.099 28 46/54

c 0.132 31 53/47

d 0.165 29 51/49

e 0.198 27 49/51

f 0.231 - 60/40

Table 5.2: The table lists the growth rates and the corresponding composition ratios of films

resulted in six experiments at different propane concentrations.

5.2.5.1 Effect of Propane Concentration on Growth Rate

A series of three complimentary experiments to both previous ones, were carried out

for investigating the effect of further increase of the C3H8 concentration on the growth

rate as well as on the film composition. The C3H8 concentration was increased in steps

from 0.066 to 0.099, 0.132, 0.165, 0.198 and 0.231 mol/m3. The SiH4 concentration

was kept constant at a value of 0.996 mol/m3 at a flow rate of 150 sccm. The growth

temperature and the pressure were kept at similar conditions to those that were estab-

lished in the first two experiments (T = 1955◦C, P = 800 mbar). The depositions were

87


carried out for 2 hours.

Figure 5.22: Diagram shows the effect of propane concentration on growth rate (Blue) and

film composition C/Si (Pink). The optimum growth rate was reached at a propane concentration

of 0.132 mol/m3. High propane concentration that exceeding a certain limit, result in high

(C/Si) ratio, which hence, lead to lower growth rates.

The growth rates of the films could be evaluated by means of SEM analysis, where the

composition ratios (C/Si) of the films were evaluated by EDX. The resulting growth

rates and the corresponding composition ratios are presented in table 5.2. The effect of

the C3H8 concentration on both, the growth rate and the film composition ratios (C/Si)

are shown in Figure 5.22. The gradual increase of C3H8 concentration to 0.132 mol/m3,

enhanced the growth rate to 31 µm/h, while further increase of C3H8 concentration to

0.165 mol/m3 and 0.198 mol/m3, resulted in slower growth rates of 29 µm/h and 27

µm/h, respectively. Further increase of C3H8 concentration to 0.231 mol/m3 led to bad

growth conditions, which resulted in the growth of polycrystalline film. In the SEM

image shown in Figure 5.23, the growth of relatively large, individual crystallites with

hexaonal surfaces could be observed.

88


Figure 5.23: SEM image shows the

growth of polycrystalline film as a result

of excessive carbon concentration caused

due to increased doses of propane.

A composition ratio of approximately C/Si:50%/50% was obtained for all the films

grown at C3H8 concentrations, which are ≤ 0.198 mol/m3, while increasing its con-

centration to 0.231 mol/m3, resulted in a higher percentage of C atoms inducing a ratio

of (C/Si:60%/40%). According to this, it can be concluded that the addition of high

C3H8 concentration is shown to have a negative effect on the growth conditions, which

limits the growth rate and leads to low Si-content. However, very low flow rates of

C3H8 are required for achieving epitaxial growth conditions and high growth rates of

SiC. This is commitment with the results of Sumakeris in [22], where low flow rate

ratio of approximately 1/8:C3H8/SiH4 was applied.

5.2.6 Growth Morphology on On/Off-oriented Surfaces

As introduced in section 2.3.3, in order to grow SiC with 4H or 6H polytypes, which

have better properties than 3C, the growth of SiC must be carried out on off-oriented

seed surfaces. This process proceeds due to the step flow nucleation mechanism lead-

ing to the homoepitaxial growth of the same seed polytype. Accordingly, the growth of

SiC was investigated on 6H-SiC seed-crystals with three different surface orientations

( on-axis, 3.5◦ off-axis and 8◦ off-axis). Three seed-crystals with those surface orien-

tations and the dimensions of 6×6×0.25 mm, were simultaneously fastened on one

89


seed-holder and hanged 30 cm below the outlet flange. The deposition was carried out

for two hours at a growth temperature of 1955◦C. The precursor flow was established

at C3H8 and SiH4 concentrations of 0.132 mol/m3 and 0.995 mol/m3 at flow rates of

20 sccm and 150 sccm, respectively.

Figure 5.24: Optical microscopic images for SiC growth on three different seed orientations

of: (a) on-axis substrate, (b) 3.5◦-off-axis, (c) 8◦-off-axis. On (b) and c) substrates, large

terraces or wavy surface morphology are visible, which reveals homo-epitaxial growth of SiC

through step-flow nucleation mechanism, while on (a), plain surface morphology is visible,

which resolves hetero-epitaxial growth of SiC through two dimensional nucleation mechanism.

The surface morphology was investigated by means of optical microscopy as seen in

Figure 5.24, which shows the different surface morphologies of the films grown on all

seed-crystals. Image a) was taken for the film grown on the on-oriented seed surface,

where a plain surface with a flat morphology could be observed on this image. In this

case, the film growth mechanism is supposed to proceed due to two dimensional nu-

cleation, which, as previously introduced, leads to heteroepitaxial growth of 3C SiC.

Image b) shows the surface morphology of the film grown on 3.5◦ off-axis surface,

which observes the growth of flat terraces ending with sharp steps. Image c) observes

the growth of a wavy surface morphology. The growth on both off-axis surfaces is

90


controlled by step-flow nucleation epitaxy. The image shown in Figure 5.25, which

was taken from a vertical cut in the 3.5◦ off-axis substrate, observes the crystal growth

of 62 µm, resulting in a growth rate of 31 µm/h. The boundary between the film and

the original seed surface is marked by a thin line as seen in this image, which seems to

be due to surface contamination with carbon during heating, since, as previously men-

tioned, at high temperatures the susceptor material decomposes to carbon, which could

be later transported by the flow and coats the seed surface. The portions contaminated

of the seed surface causes the growth of small pipes along the film depth.

Figure 5.25: Optical microscopic image

of the seed side view of the film deposited

on 3.5◦ off-axis seed. The film thickness

could be evaluated by this image to result

in a value of 62 µm. The boundary be-

tween the grown film and the original seed

surface is visible in this image, which re-

solves a seed surface contamination with

carbon.

XRD analysis was carried out for all films shown in Figure 5.24. Their resulting

diffractograms are shown in Figure 5.26. Aluminum foil was used to cover the sub-

strate side during XRD analysis in order to only analyze the film surface without the

original seed surface. The resulting X-ray diffractograms of both films grown on the

on-axis and 8◦ off-axis seeds are shown to have large similarity, except that the diffrac-

togram of the film grown on the on-axis seed included an additional reflex at θ = 42◦,

91


Figure 5.26: Three X-ray diffractograms of the films deposited on 6H-SiC seeds with different

surface orientations. On-axis: reflection observed at 42.2◦ observes the growth of 3C polytype.

This reflection was not observed in both other patterns (3.5◦ and 8◦ off-axis), which resolve that

on off-axis seeds, only 6H was grown. The reflection observed at 76◦ on both on-axis and the

8◦ off-axis seeds could not be observed on the 3.5◦ off-axis seed.

92


which indicates the growth of polytypic structure of 3C-SiC. Due to the small size of

the analyzed 3.5◦ off-axis seed and covering its side with the aluminum foil led to im-

proper scanning of its diffractogram, therefore, the XRD analysis was carried out for

this substrate without covering its side. The resulting diffractogram of the film grown

on the 3.5◦ off-axis seed included no reflections at θ > 50◦. But the reflections obtained

at θ = 34.2◦, 35.6◦, 38◦ and 45◦ were also obtained by the diffractogram obtained for

an original 6H-SiC seed as presented in section 4.2.5. Furthermore, no reflection was

found at θ = 42◦, which normally stand for observing the growth of 3C polytype. Ac-

cording to this result, it could be concluded that the deposition on the 3.5◦ off-axis seed

resulted in the homoepitaxial growth of 6H-SiC, while on on-axis surfaces the growth

of 3C-SiC will result.

5.2.7 Effect of Silane Concentration on Growth Rate.

A maximum growth rate of 31 µm/h was observed by the results presented in sec-

tion 4.2.4. This growth rate was achieved at SiH4 and C3H8 concentrations of 0.994

mol/m3 and 0.132 mol/m3 at flow rates of 150 sccm and 20 sccm, respectively. Since

further increase of the growth rate was required, the SiH4 concentration was increased

by adding 50 sccm in steps to its flow rate. Although the concentration of SiH4 was

increased, the growth rate remained limited to a value of 31 µm/h at a growth tempera-

ture of 1995◦C. Increasing the SiH4 concentration to a value of 2.240 mol/m3 at a flow

rate of 350 sccm, resulted in Si-rich film as observed by the EDX analysis.

The surface morphology of the film obtained at the highest SiH4 concentration of 2.240

mol/m3 is shown by by the image a of Figure 5.27, which observes the growth of

white Si-rich droplets with circular shapes. EDX analysis was done on on the white

93


Figure 5.27: (a): In this SEM image, the growth of white deposits on the film surface is

observed. (b): SEM image shows the location of EDX analysis, which observed that droplets

rich with Si are grown as a result of high silane concentration or low growth temperatures.

droplets and on the area around it. A composition ratio of (C/Si:18/82%) was found

for the white deposits (EDX-1), which reveals about excessive concentration of SiH4,

while on the rest of the film, a resonable composition ratio of (C/Si:52/48%) with

low percentage of Si was evaluated by (EDX-2). According to this result, it could be

considered that increasing the SiH4 concentration at this growth temperature (1955◦C)

will not result in higher growth rate, which leads to the conclusion that the growth

process is limited by the surface kinetics at such a growth temperature. Further increase

of the growth temperature seems to be necessary for increasing the surface nucleation

and consequentially consume a larger amount of the Si clusters formed in the gas-

phase.

5.2.8 Dependence of Growth Rate on Temperature

Temperature is one of the most important parameters, which plays a dominant role

in HTCVD of SiC. Increasing the reactor temperature results usually in a higher seed-

94


holder temperature (growth temperature). In HTCVD, the growth temperature plays an

important role of increasing the growth rate. Thus, as introduced in the second chapter,

the surface orientation and the polytype of the seed-crystal are decisive factors of defin-

ing the grown polytype. For example, at a high growth temperature of approximately

2000◦C, some polytypes like, 3C-SiC or 2H-SiC are not stable and tend therefore to

transform to the 6H polytype, while 4H and 6H can be grown on off-oriented surfaces

showing high stability at temperatures exceeding 2000◦C. This result was discussed in

[52]. In this section, the influence of higher growth temperatures than 1955◦C on the

growth rate and the film quality will be investigated.

A series of three experiments were carried out at different growth temperatures of

1985◦C, 2030◦C and 2060◦C. In the first experiment (at 1985◦C), three seeds with

three different orientations (on-axis, 3.5◦-off-axis and 8◦-off-axis) were simultaneously

fastened on the seed-holder and deposited. In both other experiments, the growth was

carried out on a 3.5◦-off-axis seed. The carrier gas flow rate was kept constant at 6 slm,

while the concentrations of SiH4 and C3H8 were set to 2.240 mol/m3 and 0.256 mol/m3

at flow rates of 350 sccm and 40 sccm, respectively. The seed-holder was placed at

the usual position, which is 30 cm below the outlet flange (5 cm above the ultimate

temperature point). The pressure was kept constant at 800 mbar. The depositions were

carried out for one hour in each experiment.

In the first experiment, where a growth temperature of 1985◦C was set to the seed-

holder, a maximum growth rate of 45 µm/h was obtained on the 3.5◦ off-axis seed,

while the minimal growth rate was obtained on on-axis seed, which is expected to be

caused due to the growth of the thermally instable 3C-SiC that usually results when

seeds such a surface orientation are applied at this growth temperature range. The

95


growth rate obtained on 8◦ off-axis is also less than the one obtained on the 3.5◦ off-

axis seed. According to this result, it could be concluded that the application of seeds

with 3.5◦ off-oriented surface has led to achieving the fastest growth rates at growth

temperatures above 1995◦C. Therefore, the usage of both other surface orientations

(on-axis and 8◦ off-axis) in further experiments at higher temperatures is not useful.

Tsubstrate ◦C On-Axis m/h 3.5◦ Off-Axis 8◦ Off-Axis

1955 31 31 31

1985 24 45 39

2030 74

2060 100

Table 5.3: The values of the growth rates evaluated for three films grown on different seed

orientations are listed in this table in correspondence with their growth temperature.

Further increase of the temperature to 2030◦C resulted in a growth rate enhancement

to a value of 74 µm/h, while a maximum growth rate of 100 µm/h was achieved at a

growth temperature of 2060◦C, which could be evaluated from the SEM image shown

in Figure 5.28-b. The image shows the side view of a through cut in the deposited

seed-crystal. In Figure 5.28-a, an inhomogeneous surface morphology is visible on the

substrate surface. The usual surface morphology involving terraces and steps appears

on around 60% of this image, while plain portions are visible on random areas of the

film. These areas are mostly expected to be amorphous, which seems to be caused due

to excessive concentrations of the gas-phase hydrocarbons or graphitized seed surface.

EDX analysis observed a composition ratio of (C/Si:62/38) and (C/Si:76/24) on both

seeds deposited at 2030◦C and 2060◦C, respectively. According to this, it could be con-

96


cluded that the growth temperature is strongly related with the carbon percentage of

the films. Increasing the temperature of the susceptor might lead to carbon decompo-

sition of its material, which leads to exceeding its concentration beyond the nucleation

ratio and consequentially results in unreasonable growth conditions. This problem can

be propably solved by reducing the concentration of C3H8.

Figure 5.28: (a): SEM image shows the surface morphology of the film after one hour de-

position at a temperature of 2060◦C. (b): SEM image of the seed side view. A total crystal

thickness of 350 µm was measured on this image.

In order to evaluate the total activation energy of the HTCVD process, an Arrhenius

diagram, as seen in Figure 5.29, was plotted with the results listed in Table 5.3. The

growth results obtained by Elisson in [34], are also included in the same diagram in

order to compare it with the present results. The green and the pink curves stand

for the results obtained by Elisson, whereby an activation energy of 500 kJ/mol was

evaluated. Meanwhile, an activation energy of 600 kJ/mol was evaluated for the present

study, which showed a deviation of ≈ 20% comparing it with the value obtained by

Ellison. Such a deviation seems to be caused due to geometrical or setup differences

97


between both HTCVD systems. Thus, high growth rates up to 150 µm/h were achieved

by Elisson at growth temperatures above 2150◦C as shown in the diagram, while the

maximum growth temperature achieved with our system is 2060◦C. According to both

results, it could be concluded that further increase of the growth temperature to a range

of 2060-2250◦C is necessary for achieving higher growth rates. Thus, the resulting

growth rates of the present study must not be limited by surface kinetics, but may

be due to low formation rate of the major species that require temperatures beyond

2050◦C or by the sublimation rate of the gas-phase nucleated particles.

Figure 5.29: Arrhenius plot shows the growth results obtained at different growth tempera-

tures and compared with the results obtained at the University of Linköping at a close temper-

ature range as discussed in [34].

98


5.2.9 Deposition of Thick SiC Film

In this experiment, the deposition was carried out on off-oriented seed with a tilt angle

of 3.5◦ for a period of 3 hours in order to grow thick epilayers of SiC. A growth

temperature of 1985◦C was applied. The carrier gas (helium) flow rate was set to 6

slm, while SiH4 and C3H8 concentrations were set to 2.240 mol/m3 and 0.256 mol/m3

at flow rates of 350 sccm and 40 sccm, respectively.

Figure 5.30: Two SEM images. (a): SEM image taken for the surface morphology of the

grown film after three hours of deposition deposition. (b) Magnifies the growth defects shown

in (a). Holes could be noticed on the film surface.

As shown by the SEM image of Figure 5.30, growth defects in form of holes or cracks

are visible. Such defects seem to be mainly caused by graphitization of the seed surface

during heating, particle nucleation or excessive carbon concentration in the gas-phase.

Since a ratio of (C/Si:55/45) was evaluated by means of EDX analysis on the whole

area shown by the image a) of Figure 5.30, which observes a relatively reasonable car-

bon content of the film. According to this result, it could be resolved that the concen-

99


Figure 5.31: SEM image taken for the

seed crystal fracture. A trace distinguish-

ing the original seed surface is visible in

the image. Such a defected layer is ex-

pected to be caused due to seed graphiti-

zation during heating.

tration of C3H8 was set to a proper value. The growth temperature applied is 1985◦C,

which be too low for inhibiting the particle building in the gas-phase and/or forming

the species required for the high rate of surface nucleation. Therefore, for doing such

a long experiment in future work, higher growth temperatures above 2060◦C might

reduce the particle building, which consequentially leads to enhancing the growth con-

ditions. Thus, the absence of reasonable surface treatment during heating might result

in the surface contamination with carbon causing growth defects in the initially grown

layers of the film. These defects are assumed to be enlarged during the growth of few

hundreds of micrometers. However, the film thickness could be measured using SEM

analysis. The side view of the deposited seed is shown by the SEM image of Figure

5.31, which observes a total film thickness of 150 µm that was achieved at a growth

rate of 50 µm/h. The original seed surface is still distinguished by a thin line as seen by

this image observing that the seed surface was contaminated with carbon during heat-

ing. Such a problem could be solved by further improvements of the flow geometry, or

by applying other procedures for the surface preparation during heating.

Individual crystallites with large sizes of around 1 mm width were noticed on the seed-

holder surface as seen in Figure 5.32. Such crystallites can be used as seeds for the

100

CONTENTS 5.3 Summary

growth of single crystals. Therefore, such experiments with long deposition periods,

where non-seeded growth is applied, could be exploited for achieving the growth of

large crystals that can be used as seeds for the growth of SiC by PVT or HTCVD.

Figure 5.32: The picture shows a poly-

crystalline film, which was grown on the

non-seeded area of the seed-holder. The

growth of large, individual crystallites was

noticed (marked with red circles).

5.3 Summary

The objective of this chapter was to successfully achieve homo-epitaxial growth of

6H-SiC with high growth rates and good crystal quality. In order to examine the initial

parameters of the growth process and to investigate the growth of SiC at high tempera-

tures using the HTCVD method, the non-seeded growth technique was used. Secondly,

the epitaxial growth of SiC was studied at different precursor concentrations, different

temperatures and with different orientations of the seed-crystal surface.

The vertical regions where SiC proceeds to grow, were investigated firstly. Long stripes

of graphite foil were conducted vertically in the reactor in order to indicate the verti-

cal range, where the growth of SiC occurs. The growth of polycrystalline SiC was

observed at a maximum reactor temperature of 2160◦C in a vertical region of 15-25

cm below the outlet flange. The XRD analysis proved the growth of SiC in its cubic

101

5.3 Summary CONTENTS

polytype (3C). The deposition of polycrystalline SiC was further investigated applying

a stagnant flow geometry using horizontally placed, circular seed-holders. A homoge-

neous growth of a polycrystalline SiC film could be achieved at a seed-holder position

of 15 cm below the outlet flange. Fixing the seed-crystal on a seed-holder made of

graphite foil led to a particular seed damage through a few micrometers of its adhered

side, consequentially, no mass change could be measured. Therefore, the graphite foil

was later replaced by hard, dense graphite.

In the seeded-growth experiments, 6H-SiC seeds with on-oriented surface were ini-

tially used. The growth was investigated at different growth temperatures of 1850◦C,

1950◦C and 2035◦C. Both temperatures were achieved at the same output of the MF-

Generator, while different seed-holder positions were applied. The films grown at both

growth temperatures of 1950◦C and 2035◦C included crystallites with hexagonal or

plain areas on its surface morphology, while no certain structure could be observed

within the surface morphology of the film grown at 1850◦C. This result shows the

importance of applying growth temperatures above 1900◦C, which is necessary for

improving the growth conditions targeting the epitaxial growth of SiC.

A hollow seed-holder was developed in order to reduce the contact area between the

seed and the holder. The application of such a type of seed-holder could not suppress

the usual damage of the wafer backside. Therefore in later experiments, the seed-

crystals were directly fastened on seed-holders made of dense graphite. However, the

use of a hollow seed-holder simplified the separation of the seed, which enabled the

measurement of the total seed thickness that was used for evaluating the growth rate.

A growth rate of 20 µm/h was observed by means of SEM analysis to a section of the

total seed thickness. The growth of a polycrystalline film was observed by the SEM

102


analysis of the film surface morphology. According to this, it could be resolved that

such a non-epitaxial growth of this film is a result of a low growth temperature or due

to excessive concentration of propane. Those two aspects could lead to unreasonable

gas-phase composition, slower formation rate of the species required for the epitaxial

surface nucleation or even to the building of SiC particles in the gas-phase.

In order to improve the growth parameters for achieving epitaxially grown layers of

SiC, a further increase of the temperature was carried out by shifting the seed-crystal

to lower positions towards the hot zone. A seed-holder temperature of 1995◦C was

achieved at a position of 30 cm below the outlet flange. A plain surface morphology

was observed and a maximum growth rate of 31 µm/h was achieved using an on-

axis 6H-SiC seed. The influence of the propane concentration on the growth rate was

investigated within a range of 10-40 sccm of it its flow rate, which led to the conclusion

that increasing the propane concentration beyond the nucleation value limits the growth

rate and sometimes result in the growth of polycrystalline film.

The growth on off-axis seeds was investigated in order to achieve homoepitaxial

growth of 6H-SiC. Three 6H-SiC seed-crystals with different tilt angles of 0◦, 3.5◦

and 8◦ to the basal plane (0001) were simultaneously deposited. The grown epilayers

were compared with regard to their surface morphologies and growth rates. On the

3.5◦ off-axis seed, irregular growth of wide terraces with sharp steps were observed,

while on the 8◦ off-axis seed, a wavy surface morphology was observed. A growth rate

of 31 µm/h was achieved on all surfaces. A thin line distinguishing the film from the

original seed surface was observed by means of SEM analysis. This line seems to be

caused due to the transport of the carbon decomposed of the graphite susceptor to the

seed surface.

103


For achieving higher growth rates of the SiC epilayers, further growth experiments

were carried out at higher concentrations of silane and higher growth temperatures.

The silane concentration was increased in steps by increasing its flow rate by 50 sccm

till a concentration of 2.24 mol/m3 was reached at a flow rate of 350 sccm, which re-

sulted in the deposition of Si-rich deposits at a growth temperature of 1995◦C. Thus,

the growth rate could not be enhanced by increasing the concentration of silane at

this temperature. However, increasing the growth temperature at this concentration of

Silane (2.24 mol/m3) resulted in growth rate enhancement. A maximum growth tem-

perature of 2060◦C could be achieved by our hot-wall reactor, whereat a growth rate

of 150 µm/h was reached. Indeed, the growth rate could be measured at four different

growth temperatures (1955◦C, 1980◦, 2030◦C), whereby the activation energy of the

growth process could be calculated. The resulted growth rates and the activation en-

ergy were compared with the growth results obtained in [34] using Arrhenius diagram.

104

Chapter 6

In-situ Growth Rate Measurement

6.1 In-situ Growth Rate Measurement of Polycrystalline

SiC

The growth rate was recorded by the MSB over a period of 20 minutes after an initial

heating period of 17 minutes as shown in Figure 6.1 In the initial period, the temper-

ature was increased at constant pressure of 800 mbar, which leads to lower helium

density and reduced buoyancy force. Therefore, a slight decrease of the mass was

recorded by the MSB in this period. After this period, both gaseous precursors SiH4

and C3H8 were fed into the reactor to start the SiC deposition. A constant increase of

the mass versus time was recorded in this period with a growth rate of 1.36 g/h due to

the deposition of polycrystalline SiC. The measuring technique requires a free floating

substrate (except for the link to the MSB), where the deposition of SiC, do not only

proceed on the stagnant side of the seed-holder, but (as ovserved in sec.4.1.3) also on

its back side and circumferential area. Such an undesired deposition can not be inhib-

105

6.1 In-situ Growth Rate Measurement of Polycrystalline SiC CONTENTS

ited applying such a flow geometry. For this reason the conversion of the growth from

mass per time into thickness per time is not appropriate. After 10 minutes of deposi-

tion, all flows were stopped for 6 minutes to check the response of the MSB, which led

to small fluctuations in the mass measurement during the first two minutes due to the

pressure fluctuation caused once the flows are stopped. After these two minutes, the

pressure stabilizes and no further change of the mass was recorded. Afterwards, the

deposition was continued for additional 10 minutes. As can be seen in Figure 6.1, the

increase of the mass is again linear and the growth rate is equal to the one of the first

deposition period.

Figure 6.1: The mass change is recorded versus the time over four growth periods. A heating

period is followed by a deposition period observing a growth rate of 1.38 g/h. Thereafter, the

precursor’s flow was stopped to assure a zero mass change. The deposition was continued in

the fourth interval observing an equal growth rate to the one evaluated by the first deposition

period.

106

CONTENTS 6.2 Effect of Silane Concentration on Growth Rate

6.2 Effect of Silane Concentration on Growth Rate

The growth rate dependence on precursor concentrations has been investigated. The

SiH4 concentration was gradually changed during the experiment, as seen in Figure

6.2-a. In this experiment, the pressure, the growth temperature and the flow velocity

were kept constant at 800 mbar and 1950 ◦C and 0.0075 m/s, respectively. The SiH4

concentration was initially set to 0.6683 mol/m3 (100 sccm), while the C3H8 concen-

tration was held constant at a value of 0.134 mol/m3 (20 sccm). At this concentration of

SiH4, a growth rate of 0.48 g/h was achieved, as seen in Figure 6.2-b. A further increase

of SiH4 concentration to 0.994 mol/m3 and 1.315 mol/m3, increased the growth rate

to 0.86 g/h and 0.94 g/h, respectively. Once the SiH4 concentration was increased to a

value of 1.631 mol/m3, the growth rate was nearly doubled to 1.92 g/h. Inversely, in-

creasing the SiH4 concentration to a value of 1.9414 mol/m3, resulted in lower growth

rate of 1.37 g/h, which probably resolves that the growth of polycrystalline SiC at high

SiH4 concentration could be kinetically limited within this temperature range. Thus,

the gradual increase of SiH4 concentration must not always result in a proportional in-

crease of the growth rate. The stoichiometric gas-phase composition of SiH4 and C3H8

can be calculated due to the following reaction:

3SiH4 + C3H8 = 3SiC + 16H2

According to this equation, 3 moles of SiH4 and 1 mole of C3H8 are required to pro-

duce 3 moles of SiC. Converting this on a volumetric basis, only a SiH4 flow rate of

60 sccm will be required at C3H8 flow rate of 20 sccm to fulfill the stoichiometry con-

dition, meanwhile in our case, as can be seen from the values obtained at the highest

growth rate, the concentration of SiH4 is more than 12 time the concentration of C3H8.

107

6.3 Effect of Propane Concentration on Growth Rate CONTENTS

Figure 6.2: Left: recorded mass at different silane concentrations are plotted versus time.

Right: the evaluated growth rates are plotted in correspondence with their silane concentra-

tions. The growth rate can be increased by increasing the silane concentration within a certain

range. Increasing the silane concentration beyond a certain limit results in lower growth rates.

This fact indicates the existence of a secondary carbon source in the process, which is

supposed to be supplied by the amount of carbon produced due to the evaporation of

the graphite susceptor at high temperatures.

6.3 Effect of Propane Concentration on Growth Rate

In this experiment, the SiH4 flow was held at a constant value of 100 sccm correspond-

ing to a concentration of 0.665 mol/m3, while the C3H8 concentration was gradually

changed during the experiment by increasing its flow rate in steps of 5 sccm every 10

minutes resulting in the following concentrations: 0.0335, 0.0669, 0.100 and 0.1336

mol/m3. The mass change recorded was plotted versus time as illustrated in Figure

6.3-a, where no large change in the growth rate could be observed at different C3H8

concentrations. By plotting the growth rate versus the concentration of C3H8 in the

diagram of Figure 6.3-b, a growth rate reduction was noticed at both concentrations of

C3H8 at 0.0669 and 0.100 mol/m3. However, a further increase of the C3H8 concentra-

tion from 0.1 mol/m3 to 0.1336 mol/m3, enhanced the growth rate from 0.31g/h to 0.34

108

CONTENTS 6.4 Effect of Hydrogen Concentration on Growth Rate

g/h. According to this result, it could be observed that increasing the concentration of

C3H8 could have a negative effect on the growth rate. Indeed, a further increase of the

C3H8 concentration to a value of 0.1336 mol/m3, is shown to have a better effect on

the growth rate. Thus, the highest growth rate was obtained at very low concentration

of C3H8, which is in a good agreement with the epitaxial growth results obtained in

sec.4.2.5.

Figure 6.3: Left: recorded mass at different propane concentrations are plotted versus time.

Right: the evaluated growth rates are plotted in correspondence with their propane concentra-

tions. Increasing the propane concentration within a certain range has a negative effect on the

growth rate, but a further increase of its concentration fairly enhances the growth rate.

6.4 Effect of Hydrogen Concentration on Growth Rate

The diagram illustrated in Figure 6.4 shows some similarity with the one shown in

Figure 6.3, where the growth rate decreases with increasing the concentration of C3H8.

At low concentrations of C3H8, the growth rate was measured in the first 10 minutes

without any addition of H2. A growth rate of 1.23 g/h was found for this period of

time. Afterwards, the concentration of H2 was increased from 0 to 0.3261 mol/m3 by

increasing its flow rate to 50 sccm, which caused a considerable reduction of the growth

109

6.5 Vertical Placement of the Substrate CONTENTS

rate, as seen in Figure 6.4. Descending growth rates of 0.9 g/h, 0.77 g/h and 0.65 g/h

were recorded at H2 concentrations of 0.336, 0.6471 and 0.9630 mol/m3, respectively.

By converting the growth rate of the value 0.9 g/h to thickness/time, assuming that

no deposition of SiC powder will take place on the seed-holder backside, a growth

rate of 150 µm/h will result. However, a further increase of H2 concentration to 1.274

mol/m3, resulted in a slightly enhanced growth rate, which seems to be a relatively

an obvious result due to the increased etching rates caused by adding more H2 to the

process, which decreases the hydrocarbon species in the gas-phase. However, low

concentrations of H2 seem to have a different etching effect resulting in the formation

of volatile hydrocarbons, which negatively influences the growth rate of SiC.

Figure 6.4: Left: recorded mass at different hydrogen concentrations are plotted versus time.

Right: the evaluated growth rates are plotted in correspondence with their hydrogen concen-

trations. Increasing the hydrogen concentration within a certain range has a negative effect on

the growth rate, but further increase of its concentration fairly enhances the growth rate. Such

an effect seem to be similar to the result obtained at increased propane concentrations.

6.5 Vertical Placement of the Substrate

As mentioned in the experimental part, the seed-holder was placed vertically, parallel

to the flow stream lines to enable the deposition on its both sides. In this case, ho-

110


mogeneous films are expected to be grown on both sides of the séed-holder, which is

necessary for measuring the growth rate in thickness per time. The C3H8 concentration

was varied during deposition at constant concentration of SiH4 (1.319 mol/m3). The

seed-holder mass was measured in-situ versus the time as illustrated in the diagram pre-

sented in Figure 6.5. An oscillating mass was noticed during deposition, which seems

to be the result to the new flow geometry that was established by the vertical placement

of the substrate holder. The resulting growth rates at different concentrations of C3H8

are three or four times smaller than the growth rates obtained at similar process condi-

tions where a horizontal placement of the seed-holder was applied. Due to this result

it could be concluded that placing the substrate holder in a vertical position might in-

crease the etching rate of the deposited SiC or however decreases its adsorption and

consequentially results in a slower growth rates. Indeed, the flow geometry established

by the horizontal placement of the substrate has shown a better effect on the growth

rate.

6.6 Summary

The main target of this work focused on the in-situ measurement of the growth rate of

non-seeded growth of SiC on a seed-holder. For this purpose, a MSB was successfully

integrated to the HTCVD reactor. The mass change of the deposited polycrystalline

films were accurately recorded in-situ, under different concentrations of the precursors

and for the first time at such a high growth temperature of 1950◦C. The application

of this measuring technique enabled the growth rate measurement at wide range of

the precursor concentrations that can be varied to different values in one experiment,

which saved a lot of time and experimental costs. The growth rates measured, showed

111


Figure 6.5: The growth rate is plotted versus the propane concentration, while propane and

silane concentrations are kept constant. The stepwise increase of propane concentration seems

to have a significant, negative effect on the deposition rate.

a significant dependency on the precursor concentration. Increasing the propane and

hydrogen concentrations is shown to have a negative effect on the growth rate, while

high silane concentration enhances the growth rate within a certain range.

Due to the deposition of SiC on the back and side area of the seed-holder, it was not

possible to convert the mass recorded from the mass-change/time to thickness/time.

The vertical placement of the seed-holder resulted in a growth rate reduction. There-

fore, in order to use such a system for the in-situ growth rate measurement of a seed-

crystal in an epitaxial growth process, a complete wafer of 2-inch diameter must be

fastened on the seed-holder and the internal flow geometry around the seed-holder

must be improved to inhibit the random growth of SiC powder on the seed-holder

backside. However, the successful integration of the MSB into the HTCVD system is

considered as an important step in using the gravimetric technique to study the kinetics

of the seeded growth of SiC.

112

Chapter 7

Conclusions

A vertical hot-wall reactor was constructed and successfully used for the epitaxial de-

position of SiC at very high growth temperatures. The concept of inductive heating

was applied, whereby a maximum growth temperature of 2060◦C could be achieved.

The growth temperature could be measured via an optical access using a two color

pyrometer, but only during heating where the silane concentration is still zero and the

reactor space contains no particles are usually generated when silane flows. Due to the

positioning of the inductive coil at the middle of the reactor, a low temperature zone is

established at the bottom of the susceptor and the inlet nozzle. This and the slow flow

velocity of the precursors (0.0075 m/s) prepare good conditions leading to SiC particle

generation in the gas-phase that later fall on the glass window of the optical access and

complicate the temperature measurement.

The reactor setup targeted the epitaxial growth of SiC using SiC substrates, which were

located horizontally (normal to the flow direction) in the reactor. In order to indicate

reasonable precursor concentrations and an optimal seed-holder position, a non-seeded

113

Chapter 7: Conclusions CONTENTS

growth technique was initially performed on substrates made of graphite foil. A result-

ing film of polycrystalline SiC was observed. The experimental setup (seed-holder

position and precursor concentrations) used in these experiments was then approxi-

mately applied in seeded growth experiments. Accordingly, the resulting films grown

on the SiC seeds were also polycrystalline, which indicates a significant difference in

the growth kinetics between both non-seeded and seeded growth techniques. The use

of on-axis substrates at a vertical range of 14-18 cm below the outlet flange and a pre-

cursor flow ratio of 1/3: Propane/Silane addressed some enhancement of the surface

morphology at the lowest position of the seed-holder (18 cm) where a higher growth

temperature is achieved. This leads to the conclusion that for the epitaxial growth of

SiC, a minimal growth temperature of 1950◦C is required at such reactor design and

precursor concentrations.

The use of seed-holders made of graphite foil led to the damage of the adhered side

of the seed-crystal through a depth of few micrometers. This problem was partially

solved using seed-holders made of dense, hard graphite. The use of small seed-holders

with a diameter of 50 mm was shown to have a positive effect on the flow geometry

leading to reasonable growth conditions than using seed-holders with larger diameters,

which might be caused due to that the use of large diameter seed-holders results in

parallel, long flow streams to the outer area, which is close to the perimeter of the

seed-holder surface, the flow velocity of those streams are too fast, which leads to the

conclusion that such parallel flow streams to the seed-holder surface would probably

influence the diffusion rate negatively. Thus, the use of hollow seed-holders could not

inhibit the damage of the seed-crystal (SiC wafer) backside.

114

CONTENTS Chapter 7: Conclusions

The epitaxial growth of SiC on on-axis 6H-SiC seeds could be achieved at a distance

of 30 cm below the outlet flange. This position is 5 cm above the ultimate temperature

of the susceptor walls. Upon this result, it could be resolved that such a distance (5

cm) between the seed-holder and the ultimate reactor temperature is of great impor-

tance to achieve the optimal supersaturation conditions. At a temperature of 1955◦C, a

maximum growth rate of 31 µm/h could be achieved at low concentrations of propane,

approximately 1/8:propane/silane. Increasing the concentration of the propane beyond

this ratio was observed to have a significant influence on reducing the growth rate and

leads to the growth of polycrystalline SiC at higher concentrations. According to this

result, the concentration of the hydrocarbons is considered to be playing a decisive role

of the growth kinetics of the HTCVD of SiC.

The SEM analysis of a vertical cut in the grown epilayers on on-axis SiC substrates

observed a trace separating both of the film and the original seed surface. This result

indicates that the seed surface was smudged before the growth starts, which might be

caused due to coating the seed surface with a thin film of carbon, which is expected

to be produced due to the decomposition of the graphite tube (susceptor). Such a

contamination was investigated performing different flow procedures during heating.

The lowest carbon contamination of the substrate resulted at very low pressure and

zero flow rates of the precursors and the carrier gas, where the highest contamination

resulted at helium or hydrogen flow. In order to improve the quality of the grown epi-

layers, the contamination of the seed surface must be inhibited or at least minimized.

Therefore, the research on the surface treatment must be further proceeded.

115

Chapter 7: Conclusions CONTENTS

Increasing the silane concentration to values of as high as 2.24 mol/m3 at a temperature

of 1955◦C did not result in advancing the growth rate. Si droplets were recognized by

EDX on the substrate surface. Due to this result, it could be resolved that the pro-

cess is limited by the surface kinetics, the low formation rate of the major species or

by particle formation in the gas phase at this temperature. A further increase of the

temperature to 2060◦C resulted in a significant enhancement of the growth rate to a

value of 100 µm/h on off-axis surfaces. The limited growth rates achieved on on-axis

surfaces at higher temperatures than 1955◦C, indicate the growth of 3C-SiC, which is

known for its lower stability at a temperature of 2000◦C compared with both 4H and

6H polytypes that are more stable even at higher temperatures and can be grown ho-

moepitaxially on off-axis substrates. However, it was shown that for achieving higher

growth rates up to a value of 1 mm/h, a further increase of the temperature than 2060◦C

is necessary.

The production of SiC crystal by HTCVD requires long deposition times up to 10

hours at high growth rates of up to 1 mm/h. In order to achieve optimal growth con-

ditions for obtaining such growth rates, the growth parameters involved in this process

must be optimized. The techniques used for the in-situ growth rate measurements offer

the possibility of investigating the influence of the different growth parameters on the

growth rate and saves much of preparation and experimental time. Several variations

of the growth parameters can be performed in one experiment within a short period of

time to investigate their effect on the growth rate. Therefore, a magnetic suspension

balance (MSB) was integrated to the hot-wall reactor and successfully used for the in-

116

CONTENTS Chapter 7: Conclusions

situ measurement of the growth rate. The mass change of polycrystalline films grown

directly on a seed-holder of 50 mm diameter was recorded while the precursor concen-

trations were varied over small periods of time without stopping the experiments. The

in-situ recorded mass change observed a significant dependency on the precursor con-

centrations. The growth rate was reduced by increasing one of both concentrations of

hydrogen or propane, while increasing the silane concentration resulted in high growth

rates until the concentration reached a certain limit of silane, thereafter, the growth rate

decreased at higher silane concentrations. The successful application of this technique

in such high temperatures can be considered as an important step for its application in

the epitaxial growth process of SiC using the HTCVD method.

117

Chapter 8

Appendix a: Technical Drawings

118

CONTENTS Chapter 8: Appendix a: Technical Drawings

119

Chapter 8: Appendix a: Technical Drawings CONTENTS

120


121


122


123


124


125


126


127


128


129


130


131


132


133


134

Chapter 9

Appendix b: Pictures of the Hot-wall

Reactor

Figure 9.1: Water cooled outlet flange with four symmetrical outlet openings, which can be

connected to small standard flanges of CF 40.

135

Chapter 9: Appendix b: Pictures of the Hot-wall Reactor CONTENTS

Figure 9.2: Inlet flange sealed with the quarts tube.

136

CONTENTS Chapter 9: Appendix b: Pictures of the Hot-wall Reactor

Figure 9.3: The hot-wall reactor is shown in this figure. The induction coil is fixed around

the quartz tube. The reactor connected to inlet and outlet water cooled flanges at its both ends.

137

Bibliography

[1] Peter Råback. Modeling of the sublimation growth of silicon carbide crystals.

PhD thesis, Department of Engineering Physics and Mathematics Helsinki Uni-

versity of Technology, 1999.

[2] J.A. Lely. Ber. deutsch. keram. ges. Deutsch. Keram. Ges., 32:229–231, 1955.

[3] S. N. Rashkeev W. R. L. Lambrecht, S. Limpijumnong and B. Segall. Electronic

band structure of sic polytypes: A discussion of theory. Physica Status Solidi (b),

202(1):5–33, 1977.

[4] A. Zywietz K. Karch B. Adolph K. Tenelsen F. Bechstedt, P. Käckell and J. Furth-

müller. Polytypism and properties of silicon carbide. Physica Status Solidi (b),

202(1):35–62, 1997.

[5] Hina Ashraf. Investigation of symmetries of phonons in 4H and 6H-SiC by in-

frared absorption and raman spectroscopy. PhD thesis, Linköpings Universität,

Department of Physics and Measurement Technology, 2005.

[6] Christian Janzen. Untersuchungen zur Synthese von Eisenoxid-Nanopartikeln in

der Gasphase. PhD thesis, Duisburg Uni, 9. April 2002.

138

BIBLIOGRAPHY BIBLIOGRAPHY

[7] [email protected] F. Filser & L.J. Gauckler. Materialwirtschaft 1, keramis-

che werkstoffe. ETH-Zürich, Departement Materials.

[8] V. F. Tsvetkov R. C. Glass, D. Henshall and C. H. Carter. Sic seeded crystal

growth. Physica Status Solidi (b), 202(1):147–162, 1997.

[9] I. G. Ivanov M. Syväjärvi, R. Yakimova and E. Janzén. Growth of 4h-sic from

liquid phase. Materials Science and Engineering, B46:329–332, 1997.

[10] V. E. Chelnokov M. P. Scheglov A. A. Lebedev, A. S. Tregubova and A. A.

Glagovskii. Growth and investigation of the big area lely-grown substrates. Ma-

terials Science and Engineering, B46:291–295, 1997.

[11] Yu. M. Tairov and V. F. Tsvetkov. Investigation of growth processes of ingots of

silicon carbide single crystals. Journal of Crystal Growth, 43:209–212, 1978.

[12] V. F. Tsvetkov Yu. M.Tairov. General principles of growing large-size single

crystals of various silicon carbide polytypes. J.Cryst.Growth, 52:146–150, 1981.

[13] V. F. Tsvetkov Yu. M. Tairov. Crystal growth and characterization of polytype

structures. Pergamon Press, London, page 111, 1983.

[14] D. Theis C. Weyrich G. Ziegler, P. Lang. Single crystal growth of sic substrate

material for blue light emitting diodes. Transactions on ED, ED-30(4):277–281,

1983.

[15] C. H. Carter R. F. Davis and C. E. Hunter. U.S. Patent, Re. 34,861, 1995.

[16] R.A. Stein and P.Lanig. Influence of surface energy on the growth of 6h- and

4h-sic polytypes by sublimation. Mat. Sci. Eng, B11:69, 1992.

139


[17] Yu. M. Tairov and V. F. Tsvetkov. General principles of growing largesize single

crystals of various silicon carbide polytypes. Journal of Crystal Growth, 52:146–

150, 1981.

[18] Yu. M. Tairov. Growth of bulk sic. Materials Science and Engineering, B29:83–

89, 1995.

[19] V. Balakrishna G. T. Dunne R. H. Hopkins R. N. Thomas W. A. Doolittle G. Au-

gustine, H. M. Hobgood and A. Rohatgi. High purity and semi-insulating 4h-sic

crystals grown by physical vapor transport. . Materials Science Forum, . 264-

268:. 9–12, . 1998.

[20] O. Danielssona M.K. Linnarssonc A. Henrya E. Janzen J. Zhanga, A. Ellisonb.

Epitaxial growth of 4h sic in a vertical hot-wall cvd reactor: Comparison between

up and down-flow orientations. Journal of Crystal Growth, 241:421–430, 2002.

[21] T. Kimoto. Step-controlled epitaxial growth of a-SiC and device applications.

PhD thesis, Kyoto University, Japan, 1995.

[22] J. Sumakeris H. S. Kong M. J. Paisley O. Kordina, K. Irvine and C. H. Carter.

Growth of thick epitaxial 4h-sic layers by chemical vapor deposition. Materials

Science Forum, 264-268:107–110, 1998.

[23] D.J. Larkin J.A. Powell. Process-induced morphological defects in epitaxial cvd

silicon carbide. Phys. Stat. Sol, B202:529, 1997.

[24] K. Irmscher G. Wagner. Influence of the growth conditions on the layer param-

eters of 4h-sic epilayers grown in a hot-wall reactor. Material Science Forum

Science Forum, 353-356:95, 2001.

140


[25] H. Behner A. Wiedenhofer R. Rupp, Yu. N. Makarov. Silicon carbide epitaxy

in a vertical cvd reactor: experimental results and numerical process simulation.

Physica Status Solidi (B), 202:281–304, 1997.

[26] H. D. Nordby A. A. BurkJr., M. J. O’Loughlin. Sic epitaxial layer growth in a

novel multi-wafer vapor-phase epitaxial (vpe) reactor. Journal of Crystal Growth,

Volume 200, Issues 3-4, April 1999, Pages 458-466, Volume 200, Issues 3-4:458–

466, 1999.

[27] T. Kimoto K. Fujihira and H. Matsunami. High-purity and high-quality 4h-sic

grown at high speed by chimney-type vertical hot-wall chemical vapor deposi-

tion. Appl. Phys. Lett., 80:1586, 2002.

[28] T. Jikimoto K. Izumi H. Tsuchida, I. Kamata. Growth of thick 4h-sic epilayers in

a vertical radiant-heating reactor. Mat. Res. Soc. Symp., 640:H2.12.1, 2001.

[29] T. Kimoto K. Fujihira1 and H. Matsunami. Fast epitaxial growth of high-quality

4h-sic by vertical hot-wall cvd. materials Science Forum, 433-436:161–164,

2002.

[30] A. Henry E. Janzen A. Ellison, J. Zhang. Epitaxial growth of sic in a chimney

cvd reactor. Journal of Crystal Growth, 236, Issues 1-3:225–238, 2002.

[31] A. Ellison A. S. Bakin I. G. Ivanov A. Henry R. Yakimova M. Touminen A.

Veanen E. Janzen O. Kordina, C. Hallein. High temperature chemical vapor de-

position of sic. Appl. Phys. Lett., 69:No. 10, 1996.

141


[32] B. Sundqqvist Pozina J. P. Bergman E. Janzen A. Elisson, B. Magnusson and

Veanen. Sic crystal growth by htcvd. Materials Science Forum, 457-460:9–14,

2004.

[33] Donald L. Smith. Thin-Film Deposition. 1995.

[34] J. Peterson A. Henry Q. Wahab J.P. Bergman Y.N. Makarov A. Vorob A. Vehanen

E. Janzen A. Ellison, J. Zhang. High temperature cvd growth of sic. Materials

Science and Engineering, B61-62:113–120, 1999.

[35] A. Itoh T. Kimoto and H. Matsunami. Step-controlled epitaxial growth of high-

quality sic layers. Physica Status Solidi (b), 202:247, 1997.

[36] J.T. Glass H. S. Kong and R.F. Davis. Chemical vapor deposition and charac-

terization of 6h-sic thin films on off-axis 6h-sic substrates. Journal of Applied

Physics,, 64:2672, 1988.

[37] Zhang Chuanping b Huang Ximin b Yu Xiling Yang Bailiang a, M. Isshiki a.

In-situ measurement of growth rate by laser diffraction during cdte single crystal

growth from the vapour phase. Journal of Crystal Growth, 147:399–402, 1995.

[38] O. A. Nerushev M. Jönsson and E. E. B. Campbell. In situ growth rate measure-

ments during plasma-enhanced chemical vapour deposition of vertically aligned

multiwall carbon nanotube films. Nanotechnology, 18:305702 5pp, 2007.

[39] J.R. Elliot S.A. Gokoglu M.J.Purdy S. Bammidipati, G.D. Stewart. Chemical

vapor deposition of carbon on graphite by methane pyrolysis. AIChE Journal,

42, no11:3123–3132 (25 ref.), 1996.

142


[40] A.M. Weiner S.J. Harris. Pressure and temperature effects on the kinetics and

quality of diamond films. J. Appl. Phys., 75:5026, 1994.

[41] J.C. Angus E.A. Evans. Microbalance studies of diamond nucleation and growth

rates. Diamond and related materials, 5, no 3-5 (394 p.) (18 ref.), [Notes: Part

1]:200–205, 1996.

[42] Edward A. Evans A. Salifu, G. Zhang. In situ growth rate measurements for

plasma processing of opaque materials. Thin Solid Films, 418:151–155, 2002.

[43] D. Neuschütz M. Schierling, E. Zimmermann. Deposition kinetics of al2o3 from

alcl3-co2-h2-hcl gas mixtures by thermal cvd in a hot-wall reactor. J. Phys. IV

France, 8:85–91, 1999.

[44] I.G. Ivanov O. Kordina J. Zhang C.G. Hemmingson C. Yu. Gu M.R. Leys. A. El-

lison, T. Kimoto and E. Janzen. Growth and characterisation of thick epilayers

by high temperature cvd. Material Science Forum, 264-268:103–106, 1998.

[45] J. Schoonman F.E. Kruis and B. Scarlett. Homogeneous nuculeation of silcon.

Aerosol Sci., Vol. 25, No. 7, pp. 1291 130,1, 1994, Vol. 25, No. 7:1291 130,1„

1994.

[46] Hrsg.:Verein Deutscher Ingenieure, editor. VDI-Wärmeatlas. Verein Deutscher

Ingenieure, 2006.

[47] M. Klimenkov W. Matz E. Theodossiu, H. Baumann and K. Bethge. Character-

ization of crystallinity of sic surface layers produced by ion implantation. phys.

stat. sol.(a), 182:635, 2000.

143


[48] P. Mh-tenssorf A. Ellison A. Konstantinov 0. Kordinaa E. Janzha C. Hallin,

F. Owman. In situ substrate preparation for high-quality sic chemical vapour

deposition. Journal of Crystal Growth, 18:241–253, 1997.

[49] A. Gerlitzke G. Wagner, J. Doerschel. Surface preparation of 4h-sic substrates

for hot-wall cvd of sic layers. Applied Surface Science, 184:55–59, 2001.

[50] U. Starke J. Schardt, J. Bernhardt and K. Heinz. Crystallography of the (3x3)

surface reconstruction of 3c-sic(111), 4h-sic(0001), and 6h-sic(0001) surfaces

retrieved by low-energy electron diffraction. Physical Review B, 62:15, 2000.

[51] U. Starke-K. Heinz J. Bernhardt, M. Nerding. Stable surface reconstructions on

6h-sic(0001). Materials Science and Engineering, B61-62:207–211, 1999.

[52] S. NISHINO W. S. YOO and H. MATSUNAMI. Single crystal growth of hexago-

nal sic on cubic sic by intentional polytypes. Journal of Crystal Growth, 99:278–

283, 1990.

144

Der Lebenslauf ist in der Online-Version aus Gründen des Datenschutzes nicht enthalten

145

Documents

Growth of SiC by High Temperature CVD and Application of … · 2016-05-15 · Growth of SiC by High Temperature CVD and Application of Thermo-gravimetry for an In-situ Growth Rate